WO2011093112A1 - 画像処理装置、および信号処理方法、並びにプログラム - Google Patents
画像処理装置、および信号処理方法、並びにプログラム Download PDFInfo
- Publication number
- WO2011093112A1 WO2011093112A1 PCT/JP2011/050041 JP2011050041W WO2011093112A1 WO 2011093112 A1 WO2011093112 A1 WO 2011093112A1 JP 2011050041 W JP2011050041 W JP 2011050041W WO 2011093112 A1 WO2011093112 A1 WO 2011093112A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- flicker
- image
- frame
- unit
- correction coefficient
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/70—Circuitry for compensating brightness variation in the scene
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/70—Circuitry for compensating brightness variation in the scene
- H04N23/745—Detection of flicker frequency or suppression of flicker wherein the flicker is caused by illumination, e.g. due to fluorescent tube illumination or pulsed LED illumination
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/80—Camera processing pipelines; Components thereof
- H04N23/81—Camera processing pipelines; Components thereof for suppressing or minimising disturbance in the image signal generation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/80—Camera processing pipelines; Components thereof
- H04N23/84—Camera processing pipelines; Components thereof for processing colour signals
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N25/00—Circuitry of solid-state image sensors [SSIS]; Control thereof
- H04N25/60—Noise processing, e.g. detecting, correcting, reducing or removing noise
- H04N25/62—Detection or reduction of noise due to excess charges produced by the exposure, e.g. smear, blooming, ghost image, crosstalk or leakage between pixels
Definitions
- the present invention relates to an image processing apparatus, a signal processing method, and a program. More specifically, the present invention relates to an image processing apparatus, a signal processing method, and a program that generate an image in which flicker generated in a captured image of a camera is suppressed.
- CMOS Complementary Metal Oxide Semiconductor
- flicker When an image is taken with a camera equipped with an XY address scanning type imaging device such as a CMOS (Complementary Metal Oxide Semiconductor) imaging device under illumination of a fluorescent lamp, striped luminance unevenness and color unevenness occur in the video signal. This phenomenon is called flicker. This is due to the fact that the fluorescent lamp connected to the commercial power supply (alternating current) basically repeats blinking at a cycle twice the power supply frequency and the operating principle of the image sensor.
- CMOS Complementary Metal Oxide Semiconductor
- FIG. 1 shows the following data.
- A Luminance change of fluorescent lamp
- b Schematic diagram of imaging (exposure) sequence of CMOS image sensor
- c CMOS image readout timing
- d Image of output image In FIG. 1, time (t) is from left to right Has passed.
- Fluorescent lamps are assumed to be operating in a commercial power supply area of 50 Hz.
- the fluorescent lamp repeatedly blinks at a frequency of 100 Hz which is twice the power supply frequency (50 Hz).
- the arc-shaped curve shown in (a) shows the luminance change of the fluorescent lamp.
- a luminance change is generated at a cycle of 100 Hz, that is, in units of 1/100 second.
- an image is taken with a CMOS image sensor having a rolling shutter at a frame rate of 60 frames per second.
- the exposure process is sequentially performed from the upper end row (ROW) to the lower end row (ROW) of each photographing frame as time elapses.
- an oblique dotted line is a line indicating the reset timing of the image sensor
- an oblique solid line is a line indicating the readout timing. Exposure is started after the reset indicated by the dotted line, and the period until the readout timing indicated by the solid line is the exposure time. Exposure is performed from the upper scan line of the frame toward the lower scan line by the rolling shutter.
- the area surrounded by two diagonal solid lines adjacent to each other represents one frame of the image sensor.
- a luminance change corresponding to the luminance change of the illumination occurs. That is, since the exposure timing is different for each row constituting the image frame, horizontal stripe-shaped unevenness, so-called flicker, is generated as shown in the output image of FIG.
- the output image shows four continuously shot images of frame # 1 to frame # 4. These are image frames # 1 to # 4 constituting a moving image shot at a frame rate (60 fps) of 60 frames per second.
- Frame # 3 shows an upper end portion p, an intermediate portion q, and a lower end portion r. These are shown in order to clearly show the positions corresponding to the exposure times p, q, r in the image pickup schematic diagram of the CMOS image sensor shown in FIG.
- the upper end portion p of the frame # 3 is a row portion exposed during a period when the luminance of the fluorescent lamp is bright.
- the middle portion q of frame # 3 is a row portion exposed during a period in which the luminance of the fluorescent lamp changes from a dark period to a bright period.
- the lower end r of the frame # 3 corresponds to a row portion exposed during a period when the luminance of the fluorescent lamp is dark.
- a general image sensor has a configuration in which, for example, any one of RGB wavelength lights is selectively input in units of constituent pixels of the image sensor.
- RGB array for example, a Bayer array is known.
- flicker differs for each color signal (color channel) due to the wavelength dependence of the afterglow characteristics of the phosphor of the fluorescent lamp, Color unevenness occurs due to differences in amplitude and phase.
- Patent Document 1 Japanese Patent Application Laid-Open No. 2007-174537 discloses a conventional technique that discloses a technique for preventing or suppressing such flicker.
- Patent Document 1 Japanese Patent Application Laid-Open No. 2007-174537 controls the exposure time of an image sensor by setting an electronic shutter or the like, thereby adjusting the exposure time of each row of one captured image to reduce the contrast between the rows. The processing to be performed is disclosed.
- this method places a restriction on the exposure time of the image sensor, and there is a problem that an arbitrary electronic shutter setting cannot be performed according to the shooting environment, and the degree of freedom of shooting is reduced.
- Patent Document 2 Japanese Patent Laid-Open No. 2005-347939 performs correction processing on an image signal obtained by photographing processing in order to avoid the above problem, and suppresses the influence of fluorescent lamp flicker in the video signal. Proposed configuration.
- This Patent Document 2 Japanese Patent Laid-Open No. 2005-347939 approximates a correction gain at the time of correction processing for an image signal photographed by a camera, after approximating a lighting waveform (flicker waveform) of a fluorescent lamp with a sine wave.
- the correction process is performed in accordance with the model.
- This correction process corrects the image signal in accordance with the lighting waveform (flicker waveform) of the fluorescent lamp, thereby realizing a configuration that suppresses the difference in brightness between the rows of the captured image.
- FIG. 2 shows a comparison diagram between an approximate model of a sine wave and an example of an actual luminance change of a fluorescent lamp.
- the actual fluorescent lamp lighting waveform (solid line) and the model waveform (dotted line) have different luminance changes.
- the luminance change of a general fluorescent lamp does not coincide with the approximate model of a sine wave, and in many cases, a complicated luminance change is generated as shown by a solid line in FIG.
- the effect of flicker based on the actual luminance change of the fluorescent lamp becomes more prominent when the exposure time is shortened.
- the interval between the oblique solid line and the dotted line shown in FIG. 1B is reduced, and as a result, it is more strongly affected by the luminance change of the actual fluorescent lamp.
- 3 and 4 show changes in the flicker waveform due to the difference in exposure time of the image sensor (image sensor).
- the flicker waveform is a waveform indicating luminance unevenness of each row appearing in the photographed image.
- FIG. 3 and 4 both show flicker waveforms in a video signal when shooting is performed at 60 frames / second (60 fps) by a CMOS image sensor having a rolling shutter.
- the flicker waveforms of frames # 1 to # 4 which are four consecutive frames, are shown.
- horizontal reading is the row number of the image sensor
- the vertical axis is the normalized luminance (flicker luminance) of each row.
- the normalized luminance is information for comparing relative luminance in units of lines generated by excluding the influence of the luminance of the subject.
- FIG. 3 shows flicker waveforms of frames # 1 to # 4 when the exposure time of the image sensor is 1/60 [sec].
- FIG. 4 shows flicker waveforms of frames # 1 to # 4 when the exposure time of the image sensor is 1/500 [sec].
- the flicker waveform appearing in each of the image frames # 1 to # 4 shows a gentle curve. This is a line close to a sine wave as the model waveform described above with reference to FIG.
- the flicker waveform appearing in each image frame # 1 to # 4 is different from the curve close to the sine wave shown in FIG. The line is close to the actual waveform described with reference to FIG.
- the present invention has been devised in view of such a situation, for example, and does not apply a model waveform of the luminance change of the illumination of the fluorescent lamp, but actual illumination at the time of image capturing using a captured image. Flickers are effectively generated by estimating the luminance change of the image and executing correction processing using this estimated actual waveform, specifically by setting the correction gain considering the estimated actual waveform and executing image correction processing.
- An object of the present invention is to provide an image processing apparatus, a signal processing method, and a program that can prevent or suppress the above.
- the first aspect of the present invention is: An integration processing unit that inputs image data captured by an XY address scanning type image sensor and calculates an integral value of a signal amount in units of rows; A flicker component calculation unit for detecting a flicker component included in each row of the image frame by applying an integral value of the signal amount in units of rows; A flicker correction coefficient calculation unit that calculates a flicker correction coefficient composed of a reverse phase pattern of the flicker component calculated by the flicker component calculation unit; The image processing apparatus includes a correction unit that executes a correction process using the flicker correction coefficient calculated by the flicker correction coefficient calculation unit.
- the flicker component calculating section has an integrated value of the signal amount in units of rows of the image frame to be detected as a flicker component and a frame period in which flicker having the same phase occurs.
- the flicker component detection target is obtained by comparing with the signal amount of the integrated frame, which is the signal amount in units of rows of the integrated frame obtained by adding the signal values of the corresponding pixels of the reference frame defined as a plurality of image frames included in one unit. Processing for detecting flicker components included in each row of the image frame is performed.
- the image processing apparatus further includes a still frame detection unit that detects a still frame without subject movement from successive image frames, and the flicker component calculation unit includes: The reference frame is selected from the still frames detected by the still frame detection unit.
- the flicker correction coefficient calculation unit inputs flicker phase information of an image frame to be subjected to flicker correction from the phase shift amount calculation unit, and the phase according to the flicker phase information The flicker correction coefficient corresponding to the image frame adjusted for is calculated.
- the phase shift amount calculation unit synthesizes the flicker components extracted from a plurality of frames, and estimates the flicker phase shift amount from the flicker components of one flicker period or more.
- the image processing device further selects only a flicker component having a high reliability from the flicker components calculated by the flicker component calculation unit, and the plurality of frames.
- a flicker component combining unit that combines the flicker components to generate a combined flicker component is included, and the flicker correction coefficient calculation unit calculates a flicker correction coefficient including a reverse phase pattern of the combined flicker component.
- the flicker component synthesis unit determines the reliability of the flicker component in units of frames based on the luminance information of the image frame, and has a luminance equal to or higher than a predetermined luminance threshold. Only the flicker component corresponding to the pixel having is selected as a highly reliable flicker component.
- the integration processing unit calculates an integral value of the signal amount in units of rows for each color signal (color channel), and the flicker component calculation unit Flicker components for individual signals are detected, the flicker correction coefficient calculation unit calculates flicker correction coefficients for individual color signals, and the correction unit calculates flicker correction coefficients for individual color signals calculated by the flicker correction coefficient calculation unit.
- the applied correction process is executed.
- the image processing apparatus further includes a flicker scene determination unit that determines whether or not the input image is shot under an illumination environment that generates flicker,
- the flicker scene determination unit calculates a flicker correction coefficient and executes a correction process when an image is shot in an illumination environment that generates flicker.
- the image processing apparatus combines a plurality of images with different exposure times to generate a continuous dynamic image of the same exposure time as a process before generating a wide dynamic range image.
- Correction processing to which the flicker correction coefficient is applied is executed in units of sets.
- the XY address scanning type image pickup device is a CMOS (Complementary Metal Oxide Semiconductor) image pickup device.
- the imaging apparatus includes an imaging unit and a signal processing unit that executes any one of the image processes described above.
- the third aspect of the present invention provides A signal processing method executed in an image processing apparatus, An integration processing step in which an integration processing unit inputs image data captured by an XY address scanning type imaging device and calculates an integral value of a signal amount in units of rows; A flicker component calculating unit for detecting a flicker component included in each row of the image frame by applying an integral value of the signal amount in units of rows; A flicker correction coefficient calculating unit that calculates a flicker correction coefficient composed of a reverse phase pattern of the flicker component calculated in the flicker component calculating step; In the signal processing method, the correction unit executes a correction step of executing a correction process using the flicker correction coefficient calculated in the flicker correction coefficient calculation step.
- the fourth aspect of the present invention provides A program for executing signal processing in an image processing apparatus; An integration processing step of inputting image data captured by an XY address scanning type image sensor to the integration processing unit and calculating an integral value of the signal amount in units of rows; A flicker component calculation step of causing the flicker component calculation unit to detect the flicker component included in each row of the image frame by applying an integral value of the signal amount in units of rows; A flicker correction coefficient calculation step for causing a flicker correction coefficient calculation unit to calculate a flicker correction coefficient composed of a reverse phase pattern of the flicker component calculated in the flicker component calculation step; The correction unit causes the correction unit to execute a correction process that applies the flicker correction coefficient calculated in the flicker correction coefficient calculation step.
- the program of the present invention is a program that can be provided by, for example, a storage medium or a communication medium that is provided in a computer-readable format to an information processing apparatus or a computer system that can execute various program codes.
- a program in a computer-readable format, processing corresponding to the program is realized on the information processing apparatus or the computer system.
- system is a logical set configuration of a plurality of devices, and is not limited to one in which the devices of each configuration are in the same casing.
- flickers generated in image data captured by an XY address scanning type imaging device such as a CMOS are effective in an illumination environment such as a fluorescent lamp having a change in luminance.
- the image processing apparatus calculates an integral value of the signal amount in units of rows of the correction target image to be subjected to flicker removal, and detects the flicker components included in each row of the image frame by applying this integral value. This detected flicker component is data corresponding to the flicker waveform of the actual illumination.
- a flicker correction coefficient composed of a reverse phase pattern of the flicker component is calculated, and the phase fluctuation amount of the flicker component is detected at any time and reflected in the correction coefficient, thereby executing a correction process using the flicker correction coefficient.
- correction according to the actual flicker waveform is possible, and effective flicker removal is realized.
- FIG. 10 is a diagram for describing a flicker correction coefficient calculation process executed by a flicker correction coefficient calculation unit 207.
- FIG. 10 is a diagram for describing correction processing to which a flicker correction coefficient calculated by a flicker correction coefficient calculation unit 207 is applied. It is a figure which shows the flowchart explaining an example of the sequence of the process which the image processing apparatus of this invention performs.
- the embodiment described below is an embodiment using an imaging apparatus using a CMOS image sensor as an XY address scanning type imaging device.
- the illumination environment for capturing an image is a fluorescent lamp illumination environment with a power frequency of 50 Hz
- the captured image is a processing example assuming that 60 frames / second (60 fps) video was captured.
- the present invention is not limited to the shooting conditions set as described above, and can be applied as processing for captured images in various shooting environments.
- FIG. 5 is a block diagram showing the overall configuration of the imaging apparatus 100 according to an embodiment of the present invention.
- the imaging apparatus 100 includes an imaging unit 101, a signal processing unit 102, a codec (encoding / decoding processing unit) 103, a recording / playback unit 104, a recording medium 120, a display unit 130, a control unit 150, and a memory 151.
- the imaging unit 101 is configured by a lens, an imaging element, and the like.
- the imaging device is a CMOS image sensor that is an XY address scanning type imaging device. As described above with reference to FIG. 1, the CMOS image sensor, which is an XY address scanning type image sensor, sequentially performs exposure processing from the upper row to the lower row as time elapses. The received light signal from the CMOS image sensor is photoelectrically converted and input to the signal processing unit 102 as an electrical signal.
- the signal processing unit 102 performs flicker correction processing according to the present invention in addition to processing such as white balance adjustment, gamma correction, or demosaic processing that is executed as general camera signal processing.
- processing such as white balance adjustment, gamma correction, or demosaic processing that is executed as general camera signal processing.
- the configuration of the flicker correction processing unit that executes the flicker correction process and details of the process will be described later with reference to FIG.
- Data obtained as a result of signal processing in the signal processing unit 102 is input to a codec (encoding / decoding processing unit) 103.
- the data is output to the display unit 130 and display processing is executed.
- the codec 103 executes a predetermined encoding process such as an MPEG encoding process, and outputs the processing result to the recording / reproducing unit 104.
- the recording / playback unit 104 executes a process of recording a captured image on the recording medium 120 according to a recording format defined in advance.
- the recording medium 120 is a data recordable medium such as a flash memory, a magnetic disk, and an optical disk.
- the recording / playback unit 104 When playing back an image recorded on the recording medium 120, the recording / playback unit 104 reads data from the recording medium 120, and the read data is provided to the codec 103 to perform decoding processing. Is done. An image obtained as a result of the decoding process is displayed on the display unit 130.
- the control of each of these components is executed by the control unit 150 according to a program recorded in advance in the memory 151.
- the control unit 150 has a CPU having a program execution function.
- the memory 151 is composed of RAM, ROM, and the like.
- FIG. 6 is a block diagram showing the overall configuration of the flicker correction processing unit 200 according to an embodiment of the present invention.
- the flicker correction processing unit 200 inputs a RAW image acquired by an image sensor constituting the imaging unit 101, and performs a flicker correction process on the RAW image.
- the RAW image is data immediately after output from the image sensor, and is raw data before being subjected to signal processing such as demosaic processing.
- the demosaic process is a process for setting, for example, all RGB pixel values as data corresponding to each pixel of the image sensor, and the raw data before executing the demosaic process is, for example, either RGB for each pixel of the image sensor. This is data in which only the pixel value is set.
- the flicker correction processing unit 200 illustrated in FIG. 6 receives the raw image 251 before demosaic processing, and generates a corrected image 252 in which flicker correction is performed on the raw image 251.
- this process is merely an example, and the flicker correction processing unit 200 inputs the color image on which the demosaic process has been executed, and performs the same process as described below on the color image, thereby realizing the flicker correction process. it can.
- the flicker correction processing unit 200 includes an integration processing unit 201, a flicker scene determination unit 202, a still frame detection unit 203, a phase shift amount estimation unit 204, a flicker component calculation unit 205, a flicker component synthesis unit 206, a correction coefficient calculation unit 207, and a multiplication. Part 208.
- the flicker correction processing unit 200 inputs the RAW image 251 to the integration processing unit 201.
- the integration processing unit 201 performs integration processing of signal values (pixel values) for each color channel in the horizontal direction of the RAW image 251 to obtain a one-dimensional signal amount (RP: Row Profile) in units of rows.
- the RAW data is data in which only one of RGB pixel values is set for each pixel of the image sensor.
- the integration processing unit 201 performs integration processing for each color channel of RGB, for example, in the horizontal direction of the RAW image 251 to obtain a one-dimensional signal amount (RP: Row Profile) in units of rows.
- integration processing is performed independently for each of the R, G, and B channels in units of rows, and as a one-dimensional signal amount (RP) for each of the RGB channels, Signal amount: RP_Rk (y), RP_Gk (y), RP_Bk (y) Get.
- k is the frame number of the image
- y is the vertical coordinate value of the image.
- the integration processing unit 201 performs each row (y) unit on the input RAW image frame k.
- Signal amount: RP_Rk (y), RP_Gk (y), RP_Bk (y) Is calculated.
- RGB color channel unit processing of (1) can enhance the effect of reducing color unevenness.
- the processing in the flicker correction processing unit 200 may be executed as either processing (1) or (2).
- processing (1) or (2) an example of processing performed as processing in units of RGB color channels will be described.
- RP_Rk (y), RP_Gk (y), RP_Bk (y), etc. are indicated by adding RGB, it indicates that processing is performed individually, and when RPk (y), etc., is not indicated by RGB, It is indicated that the processing is executed in common for each color channel of RGB.
- the signal amounts calculated by the integration processing unit 201: RP_Rk (y), RP_Gk (y), and RP_Bk (y) are input to the flicker scene determination unit 202.
- the flicker scene determination unit 202 performs processing (flicker determination) for determining whether or not a captured image is captured in a fluorescent light environment that generates flicker.
- the flicker scene determination unit 202 uses a one-dimensional discrete Fourier transform (DFT) to determine whether a one-dimensional signal amount RP_R, RP_G, RP_B includes a 100 Hz flicker component based on a power spectrum distribution. This processing may be performed individually for each of the RGB color channels as described above, or may be performed collectively.
- DFT discrete Fourier transform
- processing is executed on the premise that flicker is detected under the illumination of a 50 Hz fluorescent lamp.
- the illumination of the 50 Hz fluorescent lamp generates a luminance change with a period of 100 Hz as described above with reference to FIG.
- the flicker phases coincide with each other in three frame periods so that the 100 Hz component of flicker can be easily detected.
- the flicker scene determination unit 202 outputs signal amounts in units of rows of frames k ⁇ 1 to k + 1 corresponding to “same flicker generation frame period”: RP k ⁇ 1 (y), RP k ( y) and RP k + 1 (y) are calculated.
- the flicker component index value FL k (y) corresponding to the frame k is calculated according to the following equation using the signal amount in units of rows for these three consecutive frames.
- FL k (y) 3RP k (y) / (RP k ⁇ 1 (y) + RP k (y) + RP k + 1 (y))
- the flicker component index value FL k (y) calculated according to the above equation is the denominator (RP k ⁇ 1 (y) + RP k (y) + RP k + 1 (y)) of the above equation and the numerator ( 3RP k (y)) is substantially equal, and is close to 1.
- the denominator (RP k ⁇ 1 (y) + RP k (y) + RP k + 1 (y)) of the above formula and the value of the numerator (3RP k (y)) are different. The value is away from 1.
- the flicker component index value FL k (y), which is a rough flicker component, is extracted from the image according to the above formula, and then the one-dimensional discrete Fourier transform (DFT) is performed on the flicker component index value FL k (y). Perform frequency analysis with.
- the number of data elements used in the DFT is determined by cutting out data elements from the flicker component index value FL k (y) so as to correspond to one cycle of flicker, and the 0th order term ( DC) and flicker detection are performed by performing threshold processing on the ratio calculated from the power of the first order term (100 Hz).
- the above processing can be expressed by the following formula when summarized by mathematical formulas.
- the denominator corresponds to the power of the zero-order term (DC) in the frequency spectrum obtained by DFT, and the numerator in the frequency spectrum obtained by DFT. This corresponds to the power of the first order term (100 Hz).
- Re means the real part Im means the imaginary part.
- Equation 1 represents the flicker component index value FL k based on the signal amounts in units of rows of the frames k ⁇ 1 to k + 1 for the three frames described above: RP k ⁇ 1 (y), RP k (y), RP k + 1 (y). It is a calculation formula of (y).
- Formula 2 is a calculation formula showing a one-dimensional discrete Fourier transform (DFT) process for the flicker component index value FL k (y).
- Formula 3 is a formula for calculating a ratio (flicker ratio) calculated from the powers of the zero-order term (DC) and the first-order term (100 Hz) in the frequency spectrum obtained by the DFT of Formula 2. The larger the flicker ratio (PR k ) calculated by Equation 3 is, the more flicker components are included in the image frame k.
- the flicker scene determination unit 202 performs flicker scene detection by comparing the flicker ratio (PR k ) calculated by Equation 3 with a predetermined threshold value.
- the flicker ratio (PR k ) calculated by Equation 3 is equal to or greater than a predetermined flicker scene determination threshold (Tha)
- the frame k is a flicker scene, that is, an image frame causing the flicker effect. judge.
- the frame k is a flicker scene, that is, an image frame causing the influence of flicker. If the determination expression is not satisfied, the frame k is not a flicker scene, that is, the influence of flicker. It is determined that the image frame has not been generated. If it is not determined that the scene is a flicker scene, calculation of a correction coefficient for flicker removal and correction processing using the flicker correction coefficient are not performed. For a scene (frame) determined to be a flicker scene, calculation of a correction coefficient for flicker removal and correction processing using the flicker correction coefficient are executed.
- the calculation process of the flicker component index value FL k represented by the above (Equation 1) is performed not only for the frame k but also for the preceding and subsequent frames.
- the processing may be performed on the flicker components synthesized by the same method as the waveform synthesis processing based on the reliability of data described later using FL k ⁇ 1 , FL k , and FL k + 1 .
- the still frame detection unit 203 detects a frame with less movement of the subject or the camera itself in order to improve the accuracy of the correction coefficient calculated in the flicker correction coefficient calculation process executed as the process of the subsequent flicker component calculation unit 205. Perform detection.
- the still frame detection unit 203 detects a frame in which no subject moves. However, since the flicker component calculation unit 205 in the subsequent stage performs flicker component calculation using a one-dimensional signal amount that is integrated in the row direction, the still frame detection unit 203 uses the still frame based on the presence or absence of vertical motion. A determination may be made. That is, it is not necessary to perform strict stillness determination processing for lateral movement.
- the still frame detection unit 203 performs still frame detection based on the inter-frame difference value of the signal amount RP for each row calculated by the integration processing unit 201.
- a stripe pattern as flicker occurs between frames constituting a moving image. Due to the flicker movement, it may be determined that the frame is not a still frame even if the subject and the camera are stationary.
- the position of the horizontal stripes of the flicker appearing in the captured image changes in a cycle of 3 frames as described above. Therefore, if motion determination based on comparison between adjacent frames before and after is performed and still frame detection is performed, a luminance change due to flicker may be detected and erroneous processing may be performed.
- a signal amount (RP) of a plurality of consecutive frames corresponding to a frame period in which flickers having the same phase occur in this embodiment, three frames is integrated in the frame (time) direction. An RP in which the influence of flicker is mitigated is acquired, and still frame detection is performed using the signal amount RP of a plurality of frames corresponding to the flicker period.
- flicker cycle and “same flicker generation frame cycle” are defined as follows.
- Flicker cycle Flicker cycle, which is 1/100 sec in the example (100 Hz) shown in FIG.
- FIG. 7 shows consecutive image frames # (n ⁇ 1) to # (n + 3).
- the signal quantities RP n ⁇ 1 (j), RP n (j), and RP n + 1 (j) in units of rows are calculated for each of three consecutive frames from the first frame # (n ⁇ 1) to # (n + 1).
- the multi-frame integrated signal amount RP ′ n (j) is calculated according to the following equation.
- the first multi-frame integration signal amount RP ′ n obtained by integrating the RP calculated in units of 3 frames corresponding to the “same flicker generation frame period” in which flicker of the same phase occurs.
- Sk calculated from the sum of absolute values of differences for each element of the second multi-frame integral signal amount RP ′ n + 3 obtained from three consecutive frames be a still frame evaluation function.
- the still frame evaluation function Sk is calculated according to the following equation.
- the still frame evaluation function Sk calculated according to the above equation becomes a larger value as the difference between the integral signal amounts calculated in units of three frames is larger. Therefore, the value of the still frame evaluation function Sk increases for an image with motion.
- the still frame detection unit 203 detects a still frame by comparing the still frame evaluation function Sk calculated according to the above equation with a predetermined still frame determination threshold (Thb). That is, Sk ⁇ Thb When the above determination formula is satisfied, it is determined that the frame k is not a still frame. If the determination formula is not satisfied, it is determined that the frame k is a still frame.
- still frame detection process is not limited to the above method, and various existing processes such as detection by inter-frame difference can be applied as necessary.
- FIG. 8 shows the following data (a) to (c).
- (B) Multi-frame integrated signal amount in row units of 3 m frame integration frame reference signal amount (RP) (subject light component)
- FIG. 8 shows the above (a) to (c).
- (A) / (b) (c) (C) indicates that a flicker component corresponding to a row in a frame unit is calculated.
- the flicker correction coefficient applied to the flicker correction is basically a component having a phase opposite to that of the flicker component obtained here.
- the flicker correction coefficient is a correction coefficient that is multiplied by the pixel value of each row in order to remove the flicker component of each row, and is a multiplication coefficient that is composed of components that are out of phase with the flicker component.
- This correction coefficient is calculated in units of rows, and finally, the multiplication unit 208 shown in FIG. 6 performs a process of multiplying the input image in units of rows.
- the multiplication unit 208 corresponds to a correction unit that executes a correction process using a flicker correction coefficient.
- the correction image 252 from which the flicker component is removed is generated by the correction coefficient multiplication processing performed in the upper remaining portion 208.
- the row unit signal amount RP (subject light component + flicker component) in frame units is a one-dimensional signal amount (RP: Row Profile) in units of rows obtained by the integration processing unit 201.
- the 3m frame integral signal amount in units of rows reference signal amount (RP) (subject light component) is an image of three consecutive frames detected in the processing of the still frame detection unit 203 in the previous stage. It is possible to apply a multi-frame integrated signal amount RP in units of rows obtained from the integration results of Note that m is an integer of 1 or more, and more 3m frames such as 6 frames, 9 frames,... 30 frames may be used.
- RP reference signal amount
- m is an integer of 1 or more, and more 3m frames such as 6 frames, 9 frames,... 30 frames may be used.
- the reason why the integration unit is 3m frames, that is, the number of frames corresponding to a multiple of 3, is as follows.
- image capturing at 60 fps is performed under illumination with a luminance change of 100 Hz, and flicker is generated with a period of 3 frames (same flicker generation frame period). Therefore, the integrated image of the still image for 3 m frames is an image in which the flicker component is canceled.
- the flicker component is reduced, and only the light component caused by the subject is obtained.
- an integrated image for 3 m frames corresponding to a multiple of the same flicker generation frame period is used.
- the flicker component calculation unit 205 uses the integral value of the signal amount in units of rows of the image frame to be detected as the flicker component and the signal amount in units of rows of the integration frame obtained by adding the signal values of the corresponding pixels of the reference frame.
- a flicker component included in each row of an image frame as a flicker component detection target is detected by comparison with a certain amount of integrated signal of a plurality of frames.
- the reference frame is selected from the still frames detected by the still frame detection unit 203, and continuous shooting that constitutes a reference frame defined as a plurality of image frames included in one unit of a frame period in which flicker having the same phase occurs.
- a set of image frames is selected from the still frames detected by the still frame detection unit 203, and continuous shooting that constitutes a reference frame defined as a plurality of image frames included in one unit of a frame period in which flicker having the same phase occurs.
- the flicker component calculation processing executed by the flicker component calculation unit 205 is preferably performed for each color channel of the color array constituting the image sensor. That is, if the image sensor has a Bayer array, it is preferable to estimate flicker components for each of R, G, and B. This is because the afterglow characteristics of the phosphors included in the fluorescent lamp are different for each wavelength of light, and the influence on the captured image is also different for each color channel.
- the 3m frame used as the image frame for calculating the multiple-frame integrated signal amount reference signal amount (RP) applied when calculating the flicker component corresponding to the row in units of frames shown in FIG.
- RP reference signal amount
- FIG. 9 a configuration using a continuous frame (for example, 30 frames) before calculating the flicker component may be used.
- the setting shown in FIG. 9 is an example, and the setting of the frame for calculating the reference RP is a frame near the frame for calculating the flicker component, and may be a continuous frame determined as a still image.
- the flicker correction coefficient is basically a component having a phase opposite to that of the flicker component obtained here.
- a plurality of consecutive frames (three frames in this example) constituting one period of the frame period (same flicker occurrence frame period) used for estimating the flicker component is set as a reference frame.
- the flicker correction coefficient obtained for the reference frame is applied to the reference frame and other frames for correction. For example, in the example shown in FIG. 9, when a frame after frame #n is a correction target, a set of frames # n-2 to #n is a reference frame.
- the flicker correction coefficient calculated using the reference frame is applied.
- the flicker correction coefficient calculated using the reference frame is defined as a reference flicker correction coefficient.
- the multiplication unit 208 shown in FIG. 6 performs processing for multiplying the input image by the row unit by the flicker correction coefficient phase-adjusted using the reference flicker correction coefficient. By this correction coefficient multiplication processing, a corrected image 252 from which flicker components are removed is generated.
- the calculation process of the flicker correction coefficient (reference flicker correction coefficient) using the reference frame is executed as a process of the flicker component synthesis unit 206 and the flicker correction coefficient calculation unit 207 shown in FIG.
- the flicker component synthesis unit 206 calculates a flicker component corresponding to one flicker cycle with higher accuracy by using the flicker component corresponding to the reference frame.
- the flicker cycle is 1/100 sec in the example (100 Hz) shown in FIG. 1, for example, and the flicker component corresponding to the flicker pattern in one cycle unit is calculated. Processing is performed on the assumption that the flicker component for one cycle is repeatedly generated.
- the flicker correction coefficient calculation unit 207 calculates the flicker component (reference flicker correction coefficient) of the reverse phase of the flicker component by applying the flicker component generated by the flicker component synthesis unit 206 and having a high accuracy for one cycle. To do. Hereinafter, these processes will be described.
- the flicker component calculation processing of one flicker cycle with high accuracy performed by the flicker component synthesis unit 206 will be described.
- the flicker component corresponding to the frame has already been calculated by the flicker component calculation unit 205 as described with reference to FIG. Therefore, by using the flicker component corresponding to this frame, it is possible to extract the flicker component of one flicker cycle.
- the flicker component for one cycle is not simply cut out from the flicker component of the reference frame, and the robustness of the correction processing is improved as will be described below. Perform extraction.
- the high-precision flicker component calculation process executed by the flicker component synthesis unit 206 will be described in detail with reference to FIG.
- the graphs shown in FIGS. 10A, 10B, and 10C are graphs showing the flicker components of the reference frame. That is, flicker components extracted from the three consecutive image frames (frame # n-2 to frame #n) by the above-described method.
- flicker components extracted from the three consecutive image frames (frame # n-2 to frame #n) by the above-described method Even if shooting is performed with appropriate exposure, there is generally a dark region in the shooting scene where the amount of reflected light from the object is not sufficient. In such a low-luminance region, there is a high possibility that flicker components are not accurately extracted.
- a portion indicated by a dotted line is, for example, a low luminance region, and is a region where the accuracy of the flicker component is low.
- a flicker correction coefficient consisting of a coefficient having a phase opposite to the flicker component is generated from the flicker component and correction is performed, correction using a correction coefficient based on a waveform different from the original flicker waveform is performed. The image quality of the later image is degraded.
- the flicker component synthesis unit 206 calculates flicker component reliability based on luminance from each of the reference images constituting a plurality of frames in the same flicker occurrence frame period, and calculates the calculated flicker component reliability. Based on the degree, only the flicker components in the area determined to be highly reliable are selected and synthesized to generate a highly accurate synthesized flicker component (see FIG. 10D).
- the reliability of the flicker component of the reference image is, for example, a process of comparing a dark region (low luminance region) in the image with a preset luminance threshold using the above-mentioned signal amount RP (see FIG. 8) in units of rows.
- the brightness threshold value (Thc) is defined in advance, and the signal amount RP of each row of the image is compared with the brightness threshold value (Thc).
- Signal quantity RP ⁇ Thc It is determined that the flicker component extracted from a row (scan line) having sufficient luminance that satisfies the determination formula is valid. It is determined that a flicker component extracted from a low-luminance row (scan line) that does not satisfy the determination formula is invalid.
- the value of the flicker component determined to be invalid is not applied to the generation of the composite flicker component. That is, it is not used for generating the flicker correction coefficient.
- the area indicated by a dotted line is an invalid area with low reliability determined based on such a luminance level determination in units of rows.
- a region indicated by a solid line is an effective region with high reliability.
- the combined flicker component shown in FIG. 10D is generated.
- a component having a phase opposite to that of the combined flicker component is calculated as a flicker correction coefficient.
- the phase fluctuation of the flicker component due to the frequency fluctuation of the power source can be almost ignored if it is between several adjacent frames constituting the moving image.
- the flicker frequency for example, 100 Hz
- the frame rate 60 fps
- the total number of scan lines etc.
- the flicker component synthesis unit 206 applies the calculated phase shift amount of the flicker component, shifts the phase so that the respective flicker components overlap, and performs a synthesis process. That is, the flicker components in the area where reliability is low and cannot be extracted in a certain frame are combined with the phases from other frames. Even in the flicker component generated by such a synthesis process, the waveform may be discontinuous. Such a region is subjected to interpolation processing by a spline function or the like.
- the flicker component synthesis unit 206 generates a continuous synthesized flicker waveform as shown in FIG. 10D by synthesizing and interpolating the flicker components of the reference frame composed of a plurality of frames.
- the phase shift amount calculation unit 204 calculates the phase shift amount of the flicker component of the reference frame used in the flicker component synthesis unit and the phase of the flicker component of the frame to which the flicker correction coefficient is applied.
- the phase shift amount calculation unit 204 synthesizes flicker components extracted from a plurality of frames, for example, and estimates the flicker phase shift amount from the flicker components of one flicker period or more.
- the flicker correction coefficient calculation unit 207 calculates the flicker component (reference flicker correction coefficient) of the reverse phase of the flicker component by applying the flicker component for one cycle with high accuracy generated by the flicker component synthesis unit 206.
- the multiplication unit 208 shown in FIG. 6 performs the flicker correction coefficient phase-adjusted based on the phase calculated by the phase shift amount calculation unit 204 on the input image using the reference flicker correction coefficient. A process of multiplying by unit is performed. By this correction coefficient multiplication processing, a corrected image 252 from which flicker components are removed is generated.
- the phase of the flicker component changes for each captured image frame. Therefore, when performing the correction process by applying the flicker correction coefficient, it is necessary to obtain the phase information of the flicker component for each image frame and apply the flicker correction coefficient together with the phase.
- the flicker correction coefficient calculation unit 207 receives the flicker component for one cycle with high accuracy generated by the flicker component synthesis unit 206, and further updates the flicker component phase information for each reference frame sequentially updated from the phase shift amount calculation unit 204. To generate a flicker correction coefficient in units of frames combined with the phase. In other words, the flicker correction coefficient calculation unit 207 receives the flicker phase information of the image frame to be subjected to flicker correction from the phase shift amount calculation unit 204, and calculates the flicker correction coefficient corresponding to the image frame whose phase is adjusted according to the flicker phase information. Generate.
- phase shift amount of a flicker component is required.
- the calculation of the phase shift amount of the flicker component is desirably performed every fixed frame.
- the phase shift amount corresponding to the first three frames set as the same flicker occurrence frame period can be continuously used in units of three frames.
- phase shift amount estimation executed by the phase shift amount calculation unit 204 will be described below.
- the phase shift amount calculation unit 204 calculates the phase shift amount of each of the reference frames constituting the same flicker occurrence frame period unit, in this example, three frames, and provides it to the flicker correction coefficient calculation unit 207. Note that the phase shift amount calculation unit 204 calculates a phase shift amount updated by selecting a new reference frame at a preset time interval or frame interval, and provides the calculated phase shift amount to the flicker correction coefficient calculation unit 207.
- the phase shift amount calculation unit 204 performs the frequency analysis of the flicker component index value FL k indicated by (Formula 1) described above. As described above, the flicker component corresponding to the frame k using the signal amount in units of rows for a plurality of consecutive frames (three frames in this example) corresponding to the “same flicker generation frame period” constituting the reference frame.
- the phase shift amount calculation unit 204 calculates the phase shift amount of each frame constituting the reference frame by frequency analysis of the rough flicker component calculated by this (Equation 1).
- the phase shift amount calculation unit 204 performs frequency analysis by one-dimensional discrete Fourier transform (DFT) on the flicker component index value FL, similarly to the flicker detection process executed by the flicker scene determination unit 202 described above. After the flicker component index value FL is developed in the frequency domain by DFT, the second and higher harmonics are replaced with 0 in the frequency space, and then inverse DFT is performed.
- the signal amount obtained by the inverse DFT has a waveform of 100 Hz corresponding to the flicker cycle extracted from the FL.
- the phase shift amount calculation unit 204 determines the position where the waveform of 100 Hz corresponding to the flicker period is set in each frame constituting the reference frame, and calculates the phase shift amount of each frame constituting the reference frame. calculate.
- the phase shift amount calculation unit 204 provides the flicker correction coefficient calculation unit 207 with the phase shift amount of each frame constituting the reference frame calculated in this way. As described above, the phase shift amount calculation unit 204 calculates a phase shift amount by selecting and updating a new reference frame at a preset time interval or frame interval, and flicker correction coefficient calculation unit 207. To provide.
- the flicker correction coefficient calculation unit 207 receives the flicker component for one cycle with high accuracy generated by the flicker component synthesis unit 206, and further updates flicker component phase information in units of reference frames sequentially updated from the phase shift amount calculation unit 204. And a flicker correction coefficient for each frame united with the phase is generated.
- the flicker correction coefficient calculation unit 207 inputs the flicker component 301 for one cycle with high accuracy generated by the flicker component synthesis unit 206.
- This flicker component 301 corresponds to the combined flicker component shown in FIG.
- the flicker correction coefficient calculation unit 207 calculates the flicker correction coefficient of the reverse phase of the flicker component 301 for one period with high accuracy. This is the flicker correction coefficient 302 for one flicker period shown in FIG.
- each flicker correction coefficient is calculated using the phase shift amount from the composite flicker component input from the phase shift amount calculation unit 204.
- the flicker correction coefficient calculation unit 207 outputs the correction coefficient for the current frame image, and the multiplication unit 208 shown in FIG. 6 multiplies the flicker correction coefficient for each row.
- the above flicker correction processing is as described above. (1) A method of individually executing processing in units of color signal components, for example, RGB color channels; (2) A method of calculating a single signal amount RPk (y) in units of rows without distinguishing color channels (for example, RGB) and using this to perform common processing for all color channel-corresponding pixels, Any of these processes can be executed.
- the processing of the present invention is a configuration in which a flicker waveform is calculated as a waveform according to actual illumination light to obtain a correction coefficient, and is different from a configuration in which correction is performed using a model waveform such as a sine wave described as the prior art.
- a model waveform such as a sine wave described as the prior art.
- the phase shift amount may be calculated for the new reference frame at the timing when such an operation is performed.
- Power frequency 50Hz The frame rate of the image sensor is 60 fps.
- the horizontal stripe flicker generated in the video signal by the light source moves vertically in the image, in any combination of power supply frequency and frame rate of the image sensor This method is also effective.
- the luminance variation between frames may be conspicuous in the video signal after correcting the horizontal stripe flicker by this method. Since there is periodicity regarding this luminance variation, for example, if a combination of a 50 Hz power source and a 60 fps image sensor is used based on the DC value calculated for each color channel of each frame, 3 for every 3 frames. It is also possible to obtain a gain for performing correction to the median value of the DC component in the frame and perform the correction.
- FIG. 13 is a flow showing an example of the processing sequence of the flicker correction processing unit 200 shown in FIG.
- the series of processing is controlled under the control of the control unit 150 (see FIG. 5).
- step S101 signal value (pixel value) integration processing in units of rows (scan lines) is executed.
- This process is the process of the integration processing unit 201 shown in FIG.
- the integration processing unit 201 performs integration processing of signal values (pixel values) in units of horizontal scan lines (rows) of an input image (for example, the RAW image 251 shown in FIG. 5), and performs a one-dimensional signal in units of rows.
- a quantity (RP: Row Profile) is obtained.
- step S102 flicker scene determination processing is performed.
- This process is the process of the flicker scene determination unit 202 shown in FIG.
- the flicker scene determination unit 202 performs processing (flicker determination) for determining whether or not a captured image is captured in a fluorescent light environment that generates flicker.
- the flicker scene determination unit 202 uses a one-dimensional discrete Fourier transform (DFT) to determine whether a one-dimensional signal amount RP_R, RP_G, RP_B includes a 100 Hz flicker component based on a power spectrum distribution.
- DFT discrete Fourier transform
- step S103 If it is determined that the frame includes flicker, the determination in step S103 is Yes, and the processing from step S104 onward is executed. If it is determined that the frame does not include flicker, the determination in step S103 is No, the processing from step S104 onward is not executed, and the flicker correction processing ends. In this case, only other general signal processing is executed.
- step S104 still frame determination processing is performed.
- This process is the process of the still frame detecting unit 203 shown in FIG.
- the still frame detection unit 203 detects a frame with little movement of the subject or the camera itself.
- the still frame detection unit 203 performs still frame detection based on the inter-frame difference value of the signal amount RP for each row calculated by the integration processing unit 201.
- a signal amount (RP) of a plurality of continuous frames corresponding to a frame period in which flicker having the same phase occurs in this embodiment, three frames is integrated in the frame (time) direction.
- an RP in which the influence of flicker is reduced is acquired, and still frame detection is performed using the signal amount RP of a plurality of frames corresponding to the flicker period.
- the still frame evaluation function Sk is calculated, which increases as the difference between the multiple frame integrated signal amounts calculated in units of multiple frames (3 frames) in the same flicker generation frame period increases.
- the still frame is detected by comparing with a preset still frame determination threshold (Thb).
- step S105 flicker component detection processing is executed.
- This process is a process executed by the flicker component calculation unit 205 shown in FIG.
- the flicker component calculation unit 205 applies the still image frame detected in step S204 and performs a flicker component calculation process for each image frame. This process is the process described above with reference to FIGS.
- step S106 flicker component synthesis processing is executed.
- This process is the process of the flicker component synthesis unit 206 shown in FIG.
- the flicker component synthesis unit 206 calculates a flicker component corresponding to one flicker cycle with higher accuracy by using the flicker component corresponding to the reference frame.
- the flicker component reliability based on the luminance is calculated from each of the reference images constituting a plurality of frames in the same flicker occurrence frame period, and the high flicker component reliability is calculated based on the calculated flicker component reliability.
- a highly accurate synthesized flicker component is generated.
- For the phase difference of each flicker component used for synthesis a value calculated from the flicker cycle, the sensor frame rate, the total number of rows, etc. is used.
- a flicker correction coefficient calculation process is performed.
- This process is the process of the flicker correction coefficient calculation unit 207 shown in FIG.
- the flicker correction coefficient calculation unit 207 calculates the flicker correction coefficient (reference flicker correction coefficient) of the reverse phase of the flicker component by applying the flicker component for one cycle with high accuracy generated by the flicker component synthesis unit 206.
- This process is the process described with reference to FIG.
- the flicker correction coefficient calculation unit 207 receives the flicker component for one cycle with high accuracy generated by the flicker component synthesis unit 206, and further updates sequentially from the phase shift amount calculation unit 204.
- the obtained flicker component phase information of the reference frame unit is acquired, and a flicker correction coefficient of the frame unit combined with the phase is generated.
- step S108 flicker correction processing is executed.
- This process is a process in which the multiplication unit 208 multiplies the signal value (pixel value) of the image before correction by the correction unit calculated by the flicker correction coefficient calculation unit 207 shown in FIG.
- the correction coefficient is generated in units of lines, and correction is executed by multiplying the correction coefficient in units of lines by the signal value of each line of the input image.
- this processing may be performed for each color signal by calculating an independent correction coefficient for each color signal (color channel), or set for each line without distinguishing the color signal.
- a configuration in which a common correction coefficient is applied to all the row configuration pixels may be performed.
- the integration processing unit 201 calculates the integrated value of the signal amount in units of rows for each color signal (color channel), and the flicker component calculation unit 205 detects the flicker component for each color signal and flicker correction coefficient calculation unit 207. Calculates a flicker correction coefficient for each color signal, and a multiplication unit 208 as a correction unit executes a correction process using the flicker correction coefficient for each color signal calculated by the flicker correction coefficient calculation unit 207.
- the present invention analyzes the flicker component that faithfully reproduces the actual luminance change of the illumination, and executes the correction processing by the flicker correction coefficient based on this accurate flicker waveform, thereby further reducing the flicker removal processing. This can be done effectively.
- a flicker-suppressed wide dynamic range image can be generated by performing the above-described flicker correction as a process before combining images with different exposure times.
- flicker may be detected for each image frame having the same exposure time, and flicker correction processing similar to that in the above-described embodiment may be performed.
- flicker correction processing similar to that in the above-described embodiment may be performed.
- photographing is performed with a plurality of exposure times, depending on the exposure time, there may be a saturated pixel or a pixel having an extremely low value.
- the reliability of the extracted flicker component is determined based on the pixel value, and the flicker component is synthesized between the frames, thereby performing the flicker correction by setting the optimum flicker correction coefficient. It becomes possible.
- the image processing apparatus of the present invention extracts the flicker component corresponding to the illumination light at the time of image capturing from the captured image, and synthesizes the flicker waveform to thereby generate an actual light source (such as a fluorescent lamp).
- a correction coefficient corresponding to the flicker waveform is calculated, and a flicker correction process using the calculated correction coefficient is realized.
- the idea of this method is effective not only in fluorescent lamps but also in other lighting environments as long as the light source changes with time using a commercial power source.
- model approximation is not performed using a sine wave or the like, so extremely effective flicker removal correction is realized.
- the series of processes described in the specification can be executed by hardware, software, or a combined configuration of both.
- the program recording the processing sequence is installed in a memory in a computer incorporated in dedicated hardware and executed, or the program is executed on a general-purpose computer capable of executing various processing. It can be installed and run.
- the program can be recorded in advance on a recording medium.
- the program can be received via a network such as a LAN (Local Area Network) or the Internet and installed on a recording medium such as a built-in hard disk.
- the various processes described in the specification are not only executed in time series according to the description, but may be executed in parallel or individually according to the processing capability of the apparatus that executes the processes or as necessary.
- the system is a logical set configuration of a plurality of devices, and the devices of each configuration are not limited to being in the same casing.
- image data captured by an XY address scanning type imaging device such as a CMOS in an illumination environment such as a fluorescent lamp having a change in luminance is realized.
- the image processing apparatus of the present invention calculates the integral value of the signal amount in units of rows of the correction target image to be subjected to flicker removal, and applies the integral value to detect the flicker component included in each row of the image frame.
- This detected flicker component is data corresponding to the flicker waveform of the actual illumination.
- a flicker correction coefficient composed of a reverse phase pattern of the flicker component is calculated, and correction processing using the flicker correction coefficient is executed. By this processing, correction according to the actual flicker waveform is possible, and effective flicker removal is realized.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Studio Devices (AREA)
- Transforming Light Signals Into Electric Signals (AREA)
- Picture Signal Circuits (AREA)
- Color Television Image Signal Generators (AREA)
Abstract
Description
(a)蛍光灯の輝度変化
(b)CMOSイメージセンサの撮像(露光)シーケンスの模式図
(c)CMOSイメージの読み出しタイミング
(d)出力画像のイメージ
図1において時間(t)は左から右に経過している。
フレーム#3には上端部p,中間部q,下端部rをそれぞれ示している。これらは、(b)に示すCMOSイメージセンサの撮像模式図における露光時間p,q,rとの対応位置を明示するために示している。
フレーム#3の上端部pは、蛍光灯の輝度が明るい期間に露光された行部分である。
フレーム#3の中間部qは、蛍光灯の輝度が暗い期間から明るい期間に変化する期間に露光された行部分である。
フレーム#3の下端部rは、蛍光灯の輝度が暗い期間に露光された行部分に対応する。
このように各行(Row)の露光期間の蛍光灯輝度が一致しないため輝度ムラや色ムラに基づく縞模様が発生する。
この特許文献2(特開2005-347939号公報)は、蛍光灯の点灯波形(フリッカ波形)を正弦波で近似モデル化した上で、カメラの撮影した画像信号に対する補正処理時の補正ゲインを近似モデルに対応させて変更して補正処理を行うものである。この補正処理により蛍光灯の点灯波形(フリッカ波形)に応じた画像信号の補正が行われ、撮影画像の行間の明暗差を抑制する構成を実現している。
一般的な蛍光灯の輝度変化は、正弦波の近似モデルに一致することはなく、多くの場合、図2に示す実線のように複雑な輝度変化を発生させる。
図3、および図4に撮像素子(イメージセンサ)の露光時間の相違によるフリッカ波形の変化を示す。なお、ここでフリッカ波形とは撮影画像に出現する各行の輝度ムラを示す波形である。
各グラフにおいて、横読がイメージセンサの行番号であり、縦軸が各行の正規化輝度(フリッカ輝度)である。なお、正規化輝度は、被写体の輝度の影響を除外して生成される行単位の相対的な輝度の対比情報である。
図4は、イメージセンサの露光時間が短い1/500[sec]の場合のフレーム#1~#4のフリッカ波形を示している。
一方、図4に示す露光時間の短い1/500[sec]の場合、各画像フレーム#1~#4に出現するフリッカ波形は、図3に示す正弦波に近い曲線とは異なり、先に図2を参照して説明した実波形に近いラインとなっている。
XYアドレス走査型の撮像素子によって撮影された画像データを入力し、行単位の信号量の積分値を算出する積分処理部と、
前記行単位の信号量の積分値を適用して画像フレームの各行に含まれるフリッカ成分を検出するフリッカ成分算出部と、
前記フリッカ成分算出部の算出したフリッカ成分の逆相パターンからなるフリッカ補正係数を算出するフリッカ補正係数算出部と、
前記フリッカ補正係数算出部の算出したフリッカ補正係数を適用した補正処理を実行する補正部を有する画像処理装置にある。
撮像部と、上記いずれかに記載の画像処理を実行する信号処理部を備えた撮像装置にある。
画像処理装置において実行する信号処理方法であり、
積分処理部が、XYアドレス走査型の撮像素子によって撮影された画像データを入力し、行単位の信号量の積分値を算出する積分処理ステップと、
フリッカ成分算出部が、前記行単位の信号量の積分値を適用して画像フレームの各行に含まれるフリッカ成分を検出するフリッカ成分算出ステップと、
フリッカ補正係数算出部が、前記フリッカ成分算出ステップにおいて算出したフリッカ成分の逆相パターンからなるフリッカ補正係数を算出するフリッカ補正係数算出ステップと、
補正部が、前記フリッカ補正係数算出ステップにおいて算出したフリッカ補正係数を適用した補正処理を実行する補正ステップを実行する信号処理方法にある。
画像処理装置において信号処理を実行させるプログラムであり、
積分処理部に、XYアドレス走査型の撮像素子によって撮影された画像データを入力し、行単位の信号量の積分値を算出させる積分処理ステップと、
フリッカ成分算出部に、前記行単位の信号量の積分値を適用して画像フレームの各行に含まれるフリッカ成分を検出させるフリッカ成分算出ステップと、
フリッカ補正係数算出部に、前記フリッカ成分算出ステップにおいて算出したフリッカ成分の逆相パターンからなるフリッカ補正係数を算出させるフリッカ補正係数算出ステップと、
補正部に、前記フリッカ補正係数算出ステップにおいて算出したフリッカ補正係数を適用した補正処理を実行させる補正ステップを実行させるプログラムにある。
1.画像処理装置の構成例について
2.フリッカ補正処理部の構成および処理の詳細について
3.本発明の画像処理装置の処理シーケンス例について
4.広ダイナミックレンジ画像に対応する処理について
まず、本発明の画像処理装置の一実施例である撮像装置の構成例について、図5を参照して説明する。
図5は、本発明の一実施例にかかる撮像装置100の全体構成を示すブロック図である。撮像装置100は、撮像部101、信号処理部102、コーデック(符号化復号処理部)103、記録再生部104、記録メディア120、表示部130、制御部150、メモリ151を有する。
次に、信号処理部102内の一部構成として設定されるフリッカ補正処理部の構成例について図6を参照して説明する。
図6に本発明の一実施例に係るフリッカ補正処理部200の全体構成を示すブロック図を示す。
フリッカ補正処理部200は、積分処理部201、フリッカシーン判定部202、静止フレーム検出部203、位相シフト量推定部204、フリッカ成分算出部205、フリッカ成分合成部206、補正係数算出部207、乗算部208を有する。
積分処理部201は、RAW画像251の水平方向に例えばRGBの各カラーチャンネルごとに積分処理を行い、行単位の1次元の信号量(RP:Row Profile)を得る。
信号量:RP_Rk(y),RP_Gk(y),RP_Bk(y)
を得る。
なお、kは、画像のフレーム番号であり、yは画像の縦方向の座標値を表す。
積分処理部201は、このように、入力RAW画像フレームkに対して、各行(y)単位で、
信号量:RP_Rk(y),RP_Gk(y),RP_Bk(y)
を算出する。
(1)色信号成分、例えばRGBのカラーチャンネル単位で個別に処理を実行する手法と、
(2)カラーチャンネル(例えばRGB)を区別せず、行単位で1つの信号量RPk(y)を算出してこれを用いて全てのカラーチャンネル対応画素に対して共通の処理を実行する手法、
これら2つの手法がある。いずれの処理を行っても効果的なフリッカの削減効果を発揮することが可能である。なお、(1)のRGBのカラーチャンネル単位の処理は、色ムラの削減効果を高めることができる。
RP_Rk(y),RP_Gk(y),RP_Bk(y)等、RGBを付記して示す場合は、それぞれ個別に処理を行うことを示し、RPk(y)等、RGBを付記していない場合は、RGBの各カラーチャンネルに対して共通に実行される処理であることを示すものとする。
フリッカシーン判定部202は、撮影された画像がフリッカを発生する蛍光灯環境下で撮影されたかどうかの判定を行うための処理(フリッカ判定)を行う。
フリッカシーン判定部202は、1次元の離散フーリエ変換(DFT)を利用し、一次元の信号量RP_R,RP_G,RP_Bに100Hzのフリッカ成分が含まれているかをパワースペクトルの分布によって判定する。
この処理は、前述したようにRGBのカラーチャンネルごとに個別に行ってもよいし、まとめて行っても構わない。
すなわち、100Hzの輝度変化周期を持つ照明下で、60フレーム/秒(60fps)の画像撮影を行うと、
フレームn、
フレームn+1、
フレームn+2、
これら連続する画像フレームにおいて3種類のフリッカ波形が出現するが、その後は、理想的にはこの3種類のフリッカ波形が繰り返されることになる。厳密には、撮像周期のずれや電源周波数変動によってフリッカの位相がずれていく可能性があるが、近傍のフレームにおいては、この周期性を前提に処理を行うことが可能である。
次に、これらの連続する3フレーム分の行単位の信号量を用いて、フレームkに対応するフリッカ成分指標値FLk(y)を下式に従って算出する。
FLk(y)=3RPk(y)/(RPk-1(y)+RPk(y)+RPk+1(y))
一方、フリッカ成分が多い場合は、上記式の分母(RPk-1(y)+RPk(y)+RPk+1(y))と、分子(3RPk(y))の値が異なることになり、1から離れた値となる。
以上の処理を数式でまとめると、下式によって示すことができる。
Reは実数部Imは虚数部を意味する。
数式1は、前述した3フレーム分、フレームk-1~k+1の行単位の信号量:RPk-1(y)、RPk(y)、RPk+1(y)に基づくフリッカ成分指標値FLk(y)の算出式である。
数式2は、フリッカ成分指標値FLk(y)に対する1次元の離散フーリエ変換(DFT)処理を示す算出式である。
数式3は、数式2のDFTによって得られる周波数スペクトルのうち、0次項(DC)と1次項(100Hz)のパワーから計算される比率(フリッカ比率)の算出式である。
数式3によって算出されるフリッカ比率(PRk)の値が大きいほど画像フレームkにフリッカ成分がより多く含まれていることを示す。
数式3によって算出されるフリッカ比率(PRk)が予め規定したフリッカシーン判定閾値(Tha)以上である場合は、そのフレームkはフリッカシーン、すなわちフリッカの影響を発生させている画像フレームであると判定する。
PRk≧Tha
上記判定式が成立する場合、フレームkはフリッカシーン、すなわちフリッカの影響を発生させている画像フレームであると判定し、上記判定式が成立しない場合、フレームkはフリッカシーンでない、すなわちフリッカの影響を発生させていない画像フレームであると判定する。
フリッカシーンであると判定されなかった場合は、フリッカ除去のための補正係数の算出やフリッカ補正係数による補正処理は行われない。フリッカシーンであると判定されたシーン(フレーム)に対して、フリッカ除去のための補正係数の算出やフリッカ補正係数による補正処理が実行される。
フリッカの含まれる撮影シーンの場合、動画を構成するフレーム間でフリッカとしての縞模様の動きが発生する。このフリッカの動きにより、被写体、カメラが静止していても静止フレームではないと判定される可能性がある。
従って、前後の隣接フレーム同士の比較に基づく動き判別を実行して静止フレーム検出を行うと、フリッカによる輝度変化が検出され誤った処理が行われる可能性がある。
このような誤検出を避けるため、同一位相のフリッカが発生するフレーム周期に相当する複数の連続フレーム、本実施例では3フレーム分の信号量(RP)をフレーム(時間)方向に積分することによってフリッカの影響が緩和されたRPを取得して、このフリッカ周期分の複数フレームの信号量RPを用いて静止フレーム検出を行う。
(a)フリッカ周期:フリッカの周期であり、図1に示す例(100Hz)では、1/100sec
(b)同一フリッカ発生フレーム周期:同一位相のフリッカ波形の発生するフレームの周期であり、図1に示す例では3フレーム(=1/20sec)
まず、前半のフレーム#(n-1)から#(n+1)の連続する3フレームの各々について、行単位の信号量RPn-1(j)、RPn(j)、RPn+1(j)を算出し、これらの複数フレーム積分信号量RP'n(j)を下式に従って算出する。
静止フレーム検出部203は、上記式に従って算出する静止フレーム評価関数Skと、予め規定した静止フレーム判定閾値(Thb)とを比較して、静止フレームの検出を行う。すなわち、
Sk≧Thb
上記判定式が成立する場合、そのフレームkは静止フレームではないと判定する。上記判定式が成立しない場合は、フレームkは静止フレームであると判定する。
(a)フレーム単位の行単位信号量RP(被写体光成分+フリッカ成分)
(b)3mフレームの積分フレームの行単位の複数フレーム積分信号量=参照信号量(RP)(被写体光成分)
(c)フレーム単位の行対応のフリッカ成分
図8は、上記(a)~(c)について、
(a)/(b)=(c)
として、(c)フレーム単位の行対応のフリッカ成分を算出することを示している。
(a)フレーム単位の行単位信号量RP(被写体光成分+フリッカ成分)は、積分処理部201において求められる各フレーム毎の行単位の1次元の信号量(RP:Row Profile)である。
なお、mは1以上の整数であり、さらに6フレーム、9フレーム、・・・30フレーム等、より多くの3mフレームを用いてもよい。なお、積分単位を3mフレーム、すなわち3の倍数分のフレーム数とするのは以下の理由である。
このように、純粋に被写体に起因する光成分のみの行単位の1次元の信号量(RP)を算出するために、同一フリッカ発生フレーム周期の倍数に相当する3mフレーム分の積分画像を用いる。
(a)フレーム単位の行単位信号量RP(被写体光成分+フリッカ成分)
(b)3mフレームの積分フレームの行単位の複数フレーム積分信号量=参照信号量(RP)(被写体光成分)
これらのデータの演算により、
(c)フレーム単位の行対応のフリッカ成分
を算出する。
すなわち、図8に示すように、
(a)/(b)=(c)
として、(c)フレーム単位の行対応のフリッカ成分を算出する。
フリッカ成分は、イメージセンサのスキャンライン(行単位)で算出される。
なお、基準フレームは、静止フレーム検出部203の検出した静止フレームから選択され、同一位相のフリッカが発生するフレーム周期の一単位に含まれる複数の画像フレームとして定義される基準フレームを構成する連続撮影画像フレームのセットである。
例えば、図9に示すように、例えば、フリッカ成分を算出する前の連続フレーム(例えば30フレーム)を利用する構成としてもよい。
なお、図9に示す設定は一例であり、参照RPを算出するフレームの設定は、フリッカ成分を算出するフレームの近傍のフレームであり、静止画として判定された連続フレームであればよい。
フリッカ成分の推定に用いたフレーム周期(同一フリッカ発生フレーム周期)の一周期を構成する複数の連続フレーム(本例では3フレーム)を基準フレームとする。基準フレームに対して求められたフリッカ補正係数が、基準フレーム他のフレームに対して適用して補正が行われる。例えば図9に示す例では、フレーム#n以降のフレームを補正対象とする場合、フレーム#n-2~フレーム#nのセットが基準フレームとなる。
フリッカ成分合成部206は、基準フレームに対応するフリッカ成分を利用してより精度の高い1つのフリッカ周期に対応するフリッカ成分を算出する。
なお、フリッカ周期は、例えば図1に示す例(100Hz)では、1/100secであり、この1周期単位のフリッカパターンに対応するフリッカ成分を算出する。この1周期分のフリッカ成分が繰り返し発生するものとして処理を行うことになる。
フリッカ補正係数算出部207は、フリッカ成分合成部206の生成した精度の高い1周期分のフリッカ周期のフリッカ成分を適用して、フリッカ成分の逆相のフリッカ補正係数(基準フリッカ補正係数)を算出する。
以下、これらの処理について説明する。
フレーム対応のフリッカ成分は、図8を参照して説明したように、フリッカ成分算出部205において算出済みである。従って、このフレーム対応のフリッカ成分を用いれば、1つのフリッカ周期のフリッカ成分を抽出できる。
しかし、本発明の処理では、単純に基準フレームのフリッカ成分から1周期分のフリッカ成分を切り出すことはせず、以下に述べるように補正処理のロバスト性を向上させるため、より精度の高いフリッカ成分の抽出を行う。
図10(a),(b),(c)に示すグラフは、基準フレームのフリッカ成分を示すグラフである。すなわち、連続する3画像フレーム(フレーム#n-2~フレーム#n)から前述の手法により抽出されたフリッカ成分である。
しかし、適正な露出で撮影が行われたとしても、一般的に撮影シーン中には物体からの反射光量が十分ではない暗い領域がある。このような輝度の低い領域は、フリッカ成分の抽出が正確に行われない可能性が高くなる。
信号量RP≧Thc
上記判定式を満足する十分な輝度を持つ行(スキャンライン)から抽出されたフリッカ成分は有効と判定する。上記判定式を満足しない低輝度の行(スキャンライン)から抽出されたフリッカ成分は無効と判定する。無効と判定されたフリッカ成分の値は、合成フリッカ成分の生成には適用しない。すなわち、フリッカ補正係数の生成には用いないこととする。
なお、このような合成処理によって生成したフリッカ成分においても、波形が不連続となることがあり得る。このような領域についてはスプライン関数などで補間処理を施す。
先に説明したように、基準フレームを構成する「同一フリッカ発生フレーム周期」分の連続する複数フレーム(本例では3フレーム)分の行単位の信号量を用いて、フレームkに対応するフリッカ成分指標値FLk(y)は下式に従って算出される。
FLk(y)=3RPk(y)/(RPk-1(y)+RPk(y)+RPk+1(y))
・・・(数式1)
位相シフト量算出部204は、このようにして算出した基準フレームを構成する各フレームの位相シフト量をフリッカ補正係数算出部207に提供する。
なお、前述したように、位相シフト量算出部204は、予め設定した時間間隔、または、フレーム間隔で、新たな基準フレームを選択して更新した位相シフト量を算出してフリッカ補正係数算出部207に提供する。
フリッカ補正係数算出部207は、この精度の高い1周期分のフリッカ成分301の逆相のフリッカ補正係数を算出する。図11(e)に示す1つのフリッカ周期分のフリッカ補正係数302である。
この処理を実行することで、図12に示すように、(A)補正前の画像(図6の入力画像251に対応)に対する位相を合わせた補正係数の適用処理による補正が実行され。(B)補正後の画像(図6の補正画像252に対応)に示すように、フリッカを消去または減少させた画像が生成される。
(1)色信号成分、例えばRGBのカラーチャンネル単位で個別に処理を実行する手法と、
(2)カラーチャンネル(例えばRGB)を区別せず、行単位で1つの信号量RPk(y)を算出してこれを用いて全てのカラーチャンネル対応画素に対して共通の処理を実行する手法、
これらのいずれかの処理として実行することができる。
電源周波数50Hz、
イメージセンサのフレームレートが60fps、
この条件での処理例として説明を行ったが、光源によって映像信号中に発生する横縞状のフリッカが画像内を垂直方向に動く限りは、どのような電源周波数とイメージセンサのフレームレートの組み合わせにおいても本手法は有効である。
次に、本発明の画像処理装置の実行する処理のシーケンスの一例について、図13に示すフローチャートを参照して説明する。
図13に示すフローは図6に示すフリッカ補正処理部200の処理シーケンスの一例を示すフローである。なおこの一連の処理の制御は、制御部150(図5参照)の制御の下に行われる。
ステップS101において行(スキャンライン)単位の信号値(画素値)積分処理を実行する。
この処理は、図6に示す積分処理部201の処理である。積分処理部201は、入力する画像(例えば図5に示すRAW画像251)の水平方向のスキャンライン(行)単位で、信号値(画素値)の積分処理を行い、行単位の1次元の信号量(RP:Row Profile)を得る。
(1)色信号成分、例えばRGBのカラーチャンネル単位で個別に処理を実行する手法と、
(2)カラーチャンネル(例えばRGB)を区別せず、行単位で1つの信号量RPk(y)を算出してこれを用いて全てのカラーチャンネル対応画素に対して共通の処理を実行する手法、
これら2つの手法がある。
この処理は、図6に示すフリッカシーン判定部202の処理である。
フリッカシーン判定部202は、撮影された画像がフリッカを発生する蛍光灯環境下で撮影されたかどうかの判定を行うための処理(フリッカ判定)を行う。フリッカシーン判定部202は、1次元の離散フーリエ変換(DFT)を利用し、一次元の信号量RP_R,RP_G,RP_Bに100Hzのフリッカ成分が含まれているかをパワースペクトルの分布によって判定する。
前述の式(数式1)~(数式3)に従って算出されるフリッカ比率(PRk)と予め規定したフリッカシーン判定閾値(Tha)との比較により、画像フレーム単位で、フリッカを含むフレームであるか否かを判定する。
フリッカを含まないフレームであると判定した場合は、ステップS103の判定がNoとなり、ステップS104以下の処理は実行されず、フリッカ補正処理は終了する。この場合、その他の一般的な信号処理のみが実行されることになる。
この処理は、図6に示す静止レーム検出部203の処理である。静止フレーム検出部203は被写体やカメラ自体の動きが少ないフレームの検出を行う。静止フレーム検出部203は、積分処理部201の算出した行単位の信号量RPのフレーム間の差分値に基づく静止フレーム検出を行う。
図7を参照して説明したように、同一フリッカ発生フレーム周期の複数フレーム(3フレーム)単位で計算される複数フレーム積分信号量の差分が大きいほど大きな値となる静止フレーム評価関数Skを算出し、予め設定した静止フレーム判定閾値(Thb)とを比較して、静止フレームの検出を行う。
この処理は、図6に示すフリッカ成分算出部205の実行する処理である。フリッカ成分算出部205は、ステップS204において検出した静止画フレームを適用して画像フレーム単位のフリッカ成分の算出処理を行う。
この処理は、先に図8、図9を参照して説明した処理である。
(a)フレーム単位の行単位信号量RP(被写体光成分+フリッカ成分)
(b)3mフレームの積分フレームの行単位の複数フレーム積分信号量=参照信号量(RP)(被写体光成分)
(c)フレーム単位の行対応のフリッカ成分
図8は、上記(a)~(c)について、
(a)/(b)=(c)
として、(c)フレーム単位の行対応のフリッカ成分を算出する。
この処理は、図6に示すフリッカ成分合成部206の処理である。フリッカ成分合成部206は、基準フレームに対応するフリッカ成分を利用してより精度の高い1つのフリッカ周期に対応するフリッカ成分を算出する。
図10を参照して説明したように、同一フリッカ発生フレーム周期単位の複数フレームを構成する基準画像の各々から輝度に基づくフリッカ成分信頼度を算出し、算出したフリッカ成分信頼度に基づいて、高信頼度と判定された領域のフリッカ成分のみを選択して合成することで精度の高い合成フリッカ成分(図10(d)参照)を生成する。なお合成に使用する各フリッカ成分の位相差については、フリッカ周期、センサのフレームレート、総行数などから計算される値を用いる。
この処理は、図6に示すフリッカ補正係数算出部207の処理である。フリッカ補正係数算出部207は、フリッカ成分合成部206の生成した精度の高い1周期分のフリッカ成分を適用して、フリッカ成分の逆相のフリッカ補正係数(基準フリッカ補正係数)を算出する。この処理は、図11を参照して説明した処理である。
なお、この補正処理を実行する際、フリッカ補正係数算出部207は、フリッカ成分合成部206の生成した精度の高い1周期分のフリッカ成分を入力し、さらに、位相シフト量算出部204から逐次更新された基準フレーム単位のフリッカ成分位相情報を取得して、位相を併せたフレーム単位のフリッカ補正係数を生成する。
この処理は、図6に示すフリッカ補正係数算出部207の算出した補正係数を乗算部208において、補正前の画像の信号値(画素値)に乗算する処理である。補正係数は、行単位で生成されており、その行単位の補正係数を入力画像の各行の信号値に乗算して補正が実行される。
広ダイナミックレンジの画像撮影を行う処理として、露光時間の異なる複数の画像を合成する手法が知られている。
このような広ダイナミックレンジ撮像においては、イメージセンサの露光時間を周期的に変えて撮像を行い、各画素の輝度レベル等に応じた画像合成処理によって広ダイナミックレンジ画像を得る。すなわち低輝度画素部分は、長時間露光画像の画素値を適用し、高輝度画素部分は短時間露光画像の画素値を用いるといった画像合成処理によって、広ダイナミックレンジの画像を生成するものである。なお、各画素値の設定の際には、全体的な整合性を保持させるための画素値算出処理が行われる。
このような処理により、一定時間の露光時間の撮影画像では得られないダイナミックレンジの広い画像を得ることができる。
101 撮像部
102 信号処理部
103 コーデック(符号化復号処理部)
104 記録再生部
120 記録メディア
130 表示部
150 制御部
151 メモリ
200 フリッカ補正処理部
201 積分処理部
202 フリッカシーン判定部
203 静止フレーム検出部
204 位相シフト量推定部
205 フリッカ成分算出部
206 フリッカ成分合成部
207 補正係数算出部
208 乗算部
251 画像
252 補正画像
Claims (13)
- XYアドレス走査型の撮像素子によって撮影された画像データを入力し、行単位の信号量の積分値を算出する積分処理部と、
前記行単位の信号量の積分値を適用して画像フレームの各行に含まれるフリッカ成分を検出するフリッカ成分算出部と、
前記フリッカ成分算出部の算出したフリッカ成分の逆相パターンからなるフリッカ補正係数を算出するフリッカ補正係数算出部と、
前記フリッカ補正係数算出部の算出したフリッカ補正係数を適用した補正処理を実行する補正部を有する画像処理装置。 - 前記フリッカ成分算出部は、
フリッカ成分検出対象とする画像フレームの行単位の信号量の積分値と、
同一位相のフリッカが発生するフレーム周期の一単位に含まれる複数の画像フレームとして定義される基準フレームの対応画素の信号値を加算した積分フレームの行単位の信号量である複数フレーム積分信号量との比較により、フリッカ成分検出対象とする画像フレームの各行に含まれるフリッカ成分を検出する処理を行う請求項1に記載の画像処理装置。 - 前記画像処理装置は、さらに、
連続する画像フレームから被写体動きのない静止フレームを検出する静止フレーム検出部を有し、
前記フリッカ成分算出部は、
前記静止フレーム検出部の検出した静止フレームから、前記基準フレームを選択する請求項2に記載の画像処理装置。 - 前記フリッカ補正係数算出部は、
フリッカ補正対象とする画像フレームのフリッカ位相情報を位相シフト量算出部から入力し、フリッカ位相情報に応じて位相を調整した画像フレーム対応のフリッカ補正係数を算出する請求項1~3いずれかに記載の画像処理装置。 - 前記画像処理装置は、さらに、
前記フリッカ成分算出部の算出したフレーム単位のフリッカ成分から信頼度の高いフリッカ成分のみを選択し、複数フレームのフリッカ成分を合成して合成フリッカ成分を生成するフリッカ成分合成部を有し、
前記フリッカ補正係数算出部は、前記合成フリッカ成分の逆相パターンからなるフリッカ補正係数を算出する請求項1~4いずれかに記載の画像処理装置。 - 前記フリッカ成分合成部は、
フレーム単位のフリッカ成分の信頼度を画像フレームの輝度情報にもとづいて判定し、予め規定した輝度閾値以上の輝度を持つ画素対応のフリッカ成分のみを信頼度の高いフリッカ成分として選択する処理を行う請求項5に記載の画像処理装置。 - 前記積分処理部は、色信号(カラーチャンネル)個別に前記行単位の信号量の積分値を算出し、
前記フリッカ成分算出部は、前記色信号個別のフリッカ成分を検出し、
前記フリッカ補正係数算出部は、前記色信号個別のフリッカ補正係数を算出し、
前記補正部は、フリッカ補正係数算出部の算出した色信号個別のフリッカ補正係数を適用した補正処理を実行する請求項1~6いずれかに記載の画像処理装置。 - 前記画像処理装置は、さらに、
入力画像がフリッカを発生する照明環境下で撮影されたかどうかの判定を行うフリッカシーン判定部を有し、
前記フリッカシーン判定部において、画像がフリッカを発生する照明環境下で撮影された場合にフリッカ補正係数を算出して補正処理を実行する請求項1~7いずれかに記載の画像処理装置。 - 前記画像処理装置は、
複数の異なる露光時間の画像を合成して広ダイナミックレンジ画像の生成前の処理として、
同一露光時間の連続撮影画像のセット単位で、前記フリッカ補正係数を適用した補正処理を実行する請求項1~8いずれかに記載の画像処理装置。 - 前記XYアドレス走査型の撮像素子はCMOS(Complementary Metal Oxides Semiconductor)撮像素子である請求項1~9いずれかに記載の画像処理装置。
- 撮像部と、
請求項1~10いずれかに記載の画像処理を実行する信号処理部を備えた撮像装置。 - 画像処理装置において実行する信号処理方法であり、
積分処理部が、XYアドレス走査型の撮像素子によって撮影された画像データを入力し、行単位の信号量の積分値を算出する積分処理ステップと、
フリッカ成分算出部が、前記行単位の信号量の積分値を適用して画像フレームの各行に含まれるフリッカ成分を検出するフリッカ成分算出ステップと、
フリッカ補正係数算出部が、前記フリッカ成分算出ステップにおいて算出したフリッカ成分の逆相パターンからなるフリッカ補正係数を算出するフリッカ補正係数算出ステップと、
補正部が、前記フリッカ補正係数算出ステップにおいて算出したフリッカ補正係数を適用した補正処理を実行する補正ステップを実行する信号処理方法。 - 画像処理装置において信号処理を実行させるプログラムであり、
積分処理部に、XYアドレス走査型の撮像素子によって撮影された画像データを入力し、行単位の信号量の積分値を算出させる積分処理ステップと、
フリッカ成分算出部に、前記行単位の信号量の積分値を適用して画像フレームの各行に含まれるフリッカ成分を検出させるフリッカ成分算出ステップと、
フリッカ補正係数算出部に、前記フリッカ成分算出ステップにおいて算出したフリッカ成分の逆相パターンからなるフリッカ補正係数を算出させるフリッカ補正係数算出ステップと、
補正部に、前記フリッカ補正係数算出ステップにおいて算出したフリッカ補正係数を適用した補正処理を実行させる補正ステップを実行させるプログラム。
Priority Applications (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| EP11736827A EP2393284A1 (en) | 2010-01-29 | 2011-01-05 | Image processing device, signal processing device, and program |
| CN2011800014263A CN102356631A (zh) | 2010-01-29 | 2011-01-05 | 图像处理装置、信号处理方法以及程序 |
| US13/256,980 US8890975B2 (en) | 2010-01-29 | 2011-01-05 | Image processing apparatus, signal processing method and computer-readable medium for image flicker correction |
| KR1020117021398A KR20120123178A (ko) | 2010-01-29 | 2011-01-05 | 화상 처리 장치, 신호 처리 방법 및 프로그램 |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2010018714A JP5625371B2 (ja) | 2010-01-29 | 2010-01-29 | 画像処理装置、および信号処理方法、並びにプログラム |
| JP2010-018714 | 2010-01-29 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2011093112A1 true WO2011093112A1 (ja) | 2011-08-04 |
Family
ID=44319103
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2011/050041 Ceased WO2011093112A1 (ja) | 2010-01-29 | 2011-01-05 | 画像処理装置、および信号処理方法、並びにプログラム |
Country Status (6)
| Country | Link |
|---|---|
| US (1) | US8890975B2 (ja) |
| EP (1) | EP2393284A1 (ja) |
| JP (1) | JP5625371B2 (ja) |
| KR (1) | KR20120123178A (ja) |
| CN (1) | CN102356631A (ja) |
| WO (1) | WO2011093112A1 (ja) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103975578A (zh) * | 2011-12-08 | 2014-08-06 | 索尼公司 | 图像处理设备、图像处理方法以及程序 |
Families Citing this family (52)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5212170B2 (ja) * | 2009-02-25 | 2013-06-19 | 株式会社ニコン | 撮像装置 |
| CN103108132B (zh) * | 2011-11-15 | 2016-06-15 | 慧荣科技股份有限公司 | 抗闪烁摄录装置与摄录方法 |
| JP2013219708A (ja) | 2012-04-12 | 2013-10-24 | Sony Corp | 画像処理装置、および画像処理方法、並びにプログラム |
| EP2698985B1 (en) * | 2012-04-16 | 2018-05-30 | Olympus Corporation | Imaging system and imaging method |
| JP5800288B2 (ja) * | 2012-10-30 | 2015-10-28 | 株式会社デンソー | 車両用画像処理装置 |
| JP6045896B2 (ja) * | 2012-12-04 | 2016-12-14 | ハンファテクウィン株式会社Hanwha Techwin Co.,Ltd. | 評価値算出装置および評価値算出方法 |
| CN103905746B (zh) * | 2012-12-28 | 2017-02-22 | 清华大学 | 亚像素级图像偏移定位及叠加方法和装置以及视频设备 |
| JP6055681B2 (ja) * | 2013-01-10 | 2016-12-27 | 株式会社 日立産業制御ソリューションズ | 撮像装置 |
| US8934030B2 (en) * | 2013-02-20 | 2015-01-13 | Hewlett-Packard Development Company, L.P. | Suppressing flicker in digital images |
| JP6100023B2 (ja) * | 2013-02-27 | 2017-03-22 | キヤノン株式会社 | フリッカ検出装置、フリッカ補正装置、その制御方法、および制御プログラム |
| JP6116299B2 (ja) * | 2013-03-15 | 2017-04-19 | キヤノン株式会社 | 撮像装置及びその制御方法 |
| US8954277B2 (en) | 2013-05-16 | 2015-02-10 | Denso International America, Inc. | Adding visual image of traffic condition to help driver determine a route from multiple options |
| JP6060824B2 (ja) * | 2013-06-20 | 2017-01-18 | 株式会社Jvcケンウッド | 撮像装置、及びフリッカ低減方法 |
| JP6242129B2 (ja) * | 2013-09-11 | 2017-12-06 | キヤノン株式会社 | 撮像装置、その制御方法、及び制御プログラム |
| WO2015083562A1 (ja) | 2013-12-04 | 2015-06-11 | ソニー株式会社 | 画像処理装置、画像処理方法、電子機器、並びにプログラム |
| JP2015192393A (ja) * | 2014-03-28 | 2015-11-02 | ソニー株式会社 | 映像信号処理装置、映像信号処理方法およびプログラム |
| US9407832B2 (en) * | 2014-04-25 | 2016-08-02 | Himax Imaging Limited | Multi-exposure imaging system and method for eliminating rolling shutter flicker |
| US9332179B2 (en) * | 2014-06-30 | 2016-05-03 | Qualcomm Incorporated | Flash collision detection, compensation, and prevention |
| JP2016034094A (ja) * | 2014-07-31 | 2016-03-10 | ソニー株式会社 | 画像処理装置、画像処理方法、プログラム、およびイメージセンサ |
| WO2016098155A1 (ja) * | 2014-12-15 | 2016-06-23 | オリンパス株式会社 | 画像処理装置および画像処理方法 |
| CN107251543B (zh) * | 2015-02-16 | 2018-09-07 | 富士胶片株式会社 | 成像装置、闪烁检测方法及记录介质 |
| JP6304605B2 (ja) * | 2015-11-12 | 2018-04-04 | カシオ計算機株式会社 | 撮像装置、フリッカ検出方法及びプログラム |
| CN108353130A (zh) * | 2015-11-24 | 2018-07-31 | 索尼公司 | 图像处理器、图像处理方法和程序 |
| JP6691012B2 (ja) * | 2016-07-28 | 2020-04-28 | マレリ株式会社 | 車両用周辺監視装置および車両用周辺監視方法 |
| EP3503533B1 (en) * | 2016-08-16 | 2024-03-06 | Sony Group Corporation | Imaging controller, imaging control method and imaging system |
| US10972677B2 (en) * | 2016-09-14 | 2021-04-06 | Sony Corporation | Imaging control apparatus and imaging control method |
| WO2018056051A1 (ja) * | 2016-09-20 | 2018-03-29 | ソニーセミコンダクタソリューションズ株式会社 | 撮像素子および撮像素子の動作方法、撮像装置、並びに電子機器 |
| KR102743099B1 (ko) | 2017-01-13 | 2024-12-16 | 삼성전자 주식회사 | 이미지 처리 방법 및 이를 구현하는 전자 장치 |
| WO2018211511A1 (en) * | 2017-05-16 | 2018-11-22 | Technion Research & Development Foundation Limited | Computational imaging of the electric grid |
| GB2568924B (en) | 2017-11-30 | 2022-07-20 | Apical Ltd | Method of flicker reduction |
| CN110266964A (zh) * | 2018-03-12 | 2019-09-20 | 群光电子股份有限公司 | 影像撷取系统及其撷取流程调整的方法 |
| JP2020036128A (ja) | 2018-08-28 | 2020-03-05 | キヤノン株式会社 | 撮像装置及びその制御方法、プログラム、記憶媒体 |
| US11477412B2 (en) * | 2018-09-25 | 2022-10-18 | New York University | System, method, and computer-accessible medium for remote sensing of the electrical distribution grid with hypertemporal imaging |
| CN109218621B (zh) * | 2018-10-23 | 2021-03-12 | Oppo广东移动通信有限公司 | 图像处理方法、装置、存储介质及移动终端 |
| JP7336186B2 (ja) * | 2018-11-27 | 2023-08-31 | キヤノン株式会社 | 撮像装置及びその制御方法、プログラム、記憶媒体 |
| US11039082B2 (en) * | 2018-11-27 | 2021-06-15 | Canon Kabushiki Kaisha | Image capturing apparatus, control method thereof, and storage medium |
| CN114096986A (zh) | 2019-07-15 | 2022-02-25 | 谷歌有限责任公司 | 自动地分割和调整图像 |
| CN111131718B (zh) * | 2019-07-16 | 2021-05-14 | 深圳市艾为智能有限公司 | 一种具有led闪烁补偿功能的多重曝光图像融合方法和系统 |
| US12189738B2 (en) | 2019-09-09 | 2025-01-07 | Google Llc | Face authentication embedding migration and drift-compensation |
| WO2021061112A1 (en) | 2019-09-25 | 2021-04-01 | Google Llc | Gain control for face authentication |
| US10984513B1 (en) | 2019-09-30 | 2021-04-20 | Google Llc | Automatic generation of all-in-focus images with a mobile camera |
| WO2021066839A1 (en) | 2019-10-04 | 2021-04-08 | Google Llc | Autocalibration for multiple cameras using near-infrared illuminators |
| DE112019007808T5 (de) | 2019-10-10 | 2022-08-04 | Google LLC | Kamera-synchronisation und bild-markieren für eine gesichtsauthentifizierung |
| US11032486B2 (en) | 2019-10-11 | 2021-06-08 | Google Llc | Reducing a flicker effect of multiple light sources in an image |
| EP3869781A1 (en) * | 2020-02-20 | 2021-08-25 | Koninklijke Philips N.V. | Determining intensity distributions in imaging |
| CN113766114B (zh) * | 2020-06-01 | 2023-11-07 | 中移(苏州)软件技术有限公司 | 图像处理方法、装置、电子设备及存储介质 |
| CN114630054B (zh) * | 2020-12-14 | 2024-10-29 | 瑞昱半导体股份有限公司 | 具有图像闪烁检测机制的图像获取装置及其图像闪烁检测方法 |
| CN113079322B (zh) * | 2021-03-29 | 2023-04-28 | 豪威科技(上海)有限公司 | 成像系统及成像系统闪烁消除方法 |
| CN116188279A (zh) * | 2021-11-25 | 2023-05-30 | 华为技术有限公司 | 图像处理及模型训练方法及电子设备 |
| CN116744120B (zh) * | 2022-09-15 | 2024-04-12 | 荣耀终端有限公司 | 图像处理方法和电子设备 |
| US12407939B2 (en) | 2022-09-15 | 2025-09-02 | Samsung Electronics Co., Ltd. | Method of operating image sensor and image device performing the same |
| CN118521492A (zh) * | 2024-05-27 | 2024-08-20 | 京东方科技集团股份有限公司 | 图像处理方法、装置、电子设备、存储介质及程序产品 |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH11122513A (ja) * | 1997-10-09 | 1999-04-30 | Matsushita Electric Ind Co Ltd | 蛍光灯フリッカー補正装置及びそれを使用したカメラ |
| JP2001016508A (ja) * | 1999-06-30 | 2001-01-19 | Toshiba Corp | 固体撮像装置 |
| JP2004222228A (ja) * | 2002-11-18 | 2004-08-05 | Sony Corp | フリッカ低減方法、撮像装置およびフリッカ低減回路 |
| JP2005347939A (ja) | 2004-06-01 | 2005-12-15 | Sony Corp | 撮像装置および信号処理方法 |
| JP2007174537A (ja) | 2005-12-26 | 2007-07-05 | Victor Co Of Japan Ltd | 撮像装置 |
| JP2009130845A (ja) * | 2007-11-27 | 2009-06-11 | Canon Inc | フリッカ補正装置、フリッカ補正方法並びに撮像装置 |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7298401B2 (en) * | 2001-08-10 | 2007-11-20 | Micron Technology, Inc. | Method and apparatus for removing flicker from images |
| US20070024713A1 (en) * | 2005-07-28 | 2007-02-01 | Baer Richard L | Imaging parallel interface RAM |
| JP4483744B2 (ja) * | 2005-08-26 | 2010-06-16 | ソニー株式会社 | 撮像装置及び撮像制御方法 |
| JP2007097150A (ja) * | 2005-08-31 | 2007-04-12 | Pioneer Electronic Corp | 画像信号処理装置及びインターレース・プログレッシブ変換方法 |
| JP4904749B2 (ja) * | 2005-09-08 | 2012-03-28 | ソニー株式会社 | フリッカ低減方法、フリッカ低減回路及び撮像装置 |
| JP5035025B2 (ja) * | 2008-03-03 | 2012-09-26 | ソニー株式会社 | 画像処理装置、フリッカ低減方法、撮像装置及びフリッカ低減プログラム |
| JP4626689B2 (ja) * | 2008-08-26 | 2011-02-09 | ソニー株式会社 | 撮像装置、補正回路および補正方法 |
-
2010
- 2010-01-29 JP JP2010018714A patent/JP5625371B2/ja not_active Expired - Fee Related
-
2011
- 2011-01-05 CN CN2011800014263A patent/CN102356631A/zh active Pending
- 2011-01-05 WO PCT/JP2011/050041 patent/WO2011093112A1/ja not_active Ceased
- 2011-01-05 EP EP11736827A patent/EP2393284A1/en not_active Withdrawn
- 2011-01-05 US US13/256,980 patent/US8890975B2/en not_active Expired - Fee Related
- 2011-01-05 KR KR1020117021398A patent/KR20120123178A/ko not_active Withdrawn
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH11122513A (ja) * | 1997-10-09 | 1999-04-30 | Matsushita Electric Ind Co Ltd | 蛍光灯フリッカー補正装置及びそれを使用したカメラ |
| JP2001016508A (ja) * | 1999-06-30 | 2001-01-19 | Toshiba Corp | 固体撮像装置 |
| JP2004222228A (ja) * | 2002-11-18 | 2004-08-05 | Sony Corp | フリッカ低減方法、撮像装置およびフリッカ低減回路 |
| JP2005347939A (ja) | 2004-06-01 | 2005-12-15 | Sony Corp | 撮像装置および信号処理方法 |
| JP2007174537A (ja) | 2005-12-26 | 2007-07-05 | Victor Co Of Japan Ltd | 撮像装置 |
| JP2009130845A (ja) * | 2007-11-27 | 2009-06-11 | Canon Inc | フリッカ補正装置、フリッカ補正方法並びに撮像装置 |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103975578A (zh) * | 2011-12-08 | 2014-08-06 | 索尼公司 | 图像处理设备、图像处理方法以及程序 |
Also Published As
| Publication number | Publication date |
|---|---|
| JP2011160090A (ja) | 2011-08-18 |
| KR20120123178A (ko) | 2012-11-08 |
| JP5625371B2 (ja) | 2014-11-19 |
| CN102356631A (zh) | 2012-02-15 |
| US8890975B2 (en) | 2014-11-18 |
| US20120002074A1 (en) | 2012-01-05 |
| EP2393284A1 (en) | 2011-12-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP5625371B2 (ja) | 画像処理装置、および信号処理方法、並びにプログラム | |
| US20140375848A1 (en) | Image processing apparatus, image processing method, and program | |
| JP6434963B2 (ja) | 画像処理装置、画像処理方法、電子機器、並びにプログラム | |
| US9560290B2 (en) | Image processing including image correction | |
| JP6437450B2 (ja) | 画像処理装置、画像処理方法、電子機器、並びにプログラム | |
| CN101366271B (zh) | 用于校正图像中的闪烁的方法和装置 | |
| JP2013219708A (ja) | 画像処理装置、および画像処理方法、並びにプログラム | |
| US20100271515A1 (en) | Image generation apparatus and image generation method | |
| WO2017090300A1 (ja) | 画像処理装置、および画像処理方法、ならびにプログラム | |
| US10412356B2 (en) | Image processing apparatus and image processing method | |
| CN110741632B (zh) | 拍摄装置、拍摄元件以及图像处理方法 | |
| JP2014039170A (ja) | 画像処理装置、および画像処理方法、並びにプログラム | |
| WO2014020970A1 (ja) | 画像処理装置、および画像処理方法、並びにプログラム | |
| WO2017122394A1 (ja) | 撮像制御装置、および撮像装置 | |
| JP2011259390A (ja) | フリッカ検出装置 | |
| KR101004623B1 (ko) | 플래쉬 장치를 이용한 영상 품질 개선 장치 및 방법 | |
| JP4356604B2 (ja) | 映像信号処理装置、映像信号の処理方法、映像信号の処理プログラム及び映像信号の処理プログラムを記録した記録媒体 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| WWE | Wipo information: entry into national phase |
Ref document number: 201180001426.3 Country of ref document: CN |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 2011736827 Country of ref document: EP |
|
| ENP | Entry into the national phase |
Ref document number: 20117021398 Country of ref document: KR Kind code of ref document: A |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 7051/DELNP/2011 Country of ref document: IN |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 13256980 Country of ref document: US |
|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 11736827 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |