JPH0763665A - Flow-type particle image analyzer - Google Patents

Abstract

(57) [Summary] [Purpose] The same optical system is used in a plurality of measurement modes, and the particle detection unit and the feature extraction unit are combined into a single mechanism to simplify the device configuration and The present invention provides a flow-type particle image processing device that continuously captures all the images in (1) to analyze particles with high accuracy. [Structure] Sample solution 6 in which particles are suspended is mixed with flow cell 5
And the lamp 1 is always turned on to irradiate the flow cell 5, and the one-dimensional sensor 9 is provided on a plane parallel to the flow direction in the flow cell 5 and orthogonally. In order to classify the particles in the sample liquid 6 by continuously capturing the image of the sample liquid 6 passing through the image capturing area in the flow cell 5 and processing the obtained image for each pixel line. It is configured in.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a particle image analyzer, and more particularly to a flow particle image analyzer for automatic analysis of blood cell components, urine sediment and the like. In particular, in a flow cytometer, it is used for analysis of particle images of cells flowing in the flow cell.

[0002]

2. Description of the Related Art In a conventional particle image analyzer, in order to classify and analyze cells in blood or cells and particles present in urine, a sample is prepared on a slide glass and observed with a microscope. I've been told. In the case of urine, since the particle concentration in urine is low, the sample was observed by preliminarily concentrating it in a centrifuge.

As an apparatus for automating these observation and inspection operations, a sample such as blood is spread on a slide glass, set on a microscope, and the microscope stage is automatically scanned to detect the presence of particles. The microscopic stage is stopped at the position where the image is taken, a still image of the particle is photographed, and the feature extraction by the image processing technology and the pattern recognition method are used to classify the particle present in the sample. .

However, in the above method, it takes time to prepare a sample, and it is necessary to find particles while mechanically moving the microscope stage and move the particles to an appropriate image capturing area. Therefore, there are drawbacks that analysis takes time and the mechanism becomes complicated.

Therefore, a flow cytometer method has been proposed in which the smear sample as described above is not prepared, and the sample sample is allowed to flow in the flow cell while being suspended in a liquid and optically analyzed. There is. The method using this flow cytometer is
It observes the fluorescence intensity and scattered light intensity from each particle in the sample, and has a processing capacity of several thousand particles per second.

However, this method is limited to classification by particle size, and it is difficult to observe the feature quantity reflecting the morphological characteristics, and the conventional particle classification by morphological characteristics performed under the microscope is performed. Can not do it.
Incidentally, in the flow cytometer method or the particle analysis apparatus which does not take a still image, as an example in which a particle detection unit is separately provided, there are JP-A-60-260830 and JP-A-63-2.
Techniques described in Japanese Patent No. 31244, Japanese Patent Laid-Open No. 1-245131, etc. are known.

Further, an attempt has been proposed in which a particle image in a continuously flowing sample sample is photographed and particles are analyzed / classified from individual particle images. As for those related to this, JP-A-57-500995 and JP-A-6-500995
The technique described in Japanese Patent Laid-Open No. 3-94156 is known. Among them, in the technique described in Japanese Patent Publication No. 57-500995, the sample sample is passed through a channel having a special shape, in which particles in the sample are flowed into a wide imaging region, and a still image is taken by a flash lamp. Particle analysis is performed.

In this case, when a magnified image of sample particles is projected onto the CCD camera using a microscope, a flash lamp, which is a pulse light source, periodically emits light in synchronization with the operation of the CCD camera. Since the light emission time of the pulse light source is short, a still image can be obtained even when particles are continuously flowing, and the CCD camera can take 30 still images per second.

In the technique disclosed in Japanese Patent Laid-Open No. 63-94156, an optical system for particle detection is provided upstream of the particle image capturing area in the sample flow, in addition to the still image capturing system. The particle detection unit knows the particle passage in advance, and when the particle reaches the particle image capturing area, the flash lamp is turned on at an appropriate timing.

According to this technique, it is possible to detect the passage of particles and shoot a still image at a timing only when the passage of particles is detected without periodically emitting light from a pulsed light source. Even in the case of thin sample specimens, it does not process meaningless images in the absence of particles.

Next, in the conventional particle image analysis apparatus, in order to improve the measurement accuracy, the particles are observed by changing the magnification of the particle image, and the type of particles to be analyzed is limited, but the volume of the measurement sample is limited. In some cases, a plurality of measurement modes are provided for one measurement sample, such as increasing the number of measurements.

For example, in the case of measuring particles of a small size, the measurement volume of a sample is small, but in a measurement mode in which the image magnification is high for particles, and in the case of measuring a small number of particles of a large size, the image magnification is set for particles. A measurement mode is provided in which the image capturing area is low and the volume of the measurement sample is large. In the latter case, in order to increase the sample volume, the sample thickness in the optical axis direction of the passing light is also increased at the same time.

[0013]

In the conventional flow-type particle image analysis apparatus of each of the above-mentioned methods, in the flow cytometer method in which a still image is not captured, the classification is limited to the size of the particles and the form of the particles is limited. It is difficult to observe the biological features, and there is a problem that the particles cannot be classified by the morphological features performed under the microscope.

Further, a still image of the particles which are continuously flowing as described above is imaged by using a television camera as an image pickup means, and the image is analyzed to detect a plurality of kinds of particles contained in the sample liquid. The flow-type particle image analyzer for classifying and counting the number has the following problems. That is, it takes a finite time to capture a particle image with a television camera from the time when the particle is detected by the particle detection system, convert the particle image into an electrical signal proportional to the grayscale image of the particle, and transfer the electrical signal as a video signal.

Image transfer work continues during this finite time, but the flash lamp is not lit so as not to disturb the imaged image during that time, and until the transfer work ends, the following new Even if the particles pass through, the image cannot be taken, and the information in the meantime must be discarded. In this way, there is a dead time (hereinafter referred to as dead time) in which an image cannot be captured, and particles that have passed during this dead time are not captured and image processing cannot be performed.

In the sample solution having a low concentration, the probability that particles that have arrived at the image capturing area following the previous particles will be present during the dead time is small, but in the sample solution having a high particle concentration, the probability is high. Since the dead time occupies most of the measurement time, it is difficult to accurately classify multiple types of particles contained in the sample liquid from the number of image-processed particles and to know the number. There was also a problem.

In order to improve the following measurement accuracy, the flow-type particle image analyzer having a plurality of measurement modes requires a means for switching the image magnification in the plurality of measurement modes. Further, when the image magnification is changed, the amount of light on the image pickup surface of the image pickup means, for example, a TV camera is changed accordingly, and the magnitude of the image output signal is changed.

Since the TV camera usually has a means for setting a gain control in order to cope with the change in the light quantity on the image pickup surface, the gain control is provided for each measurement mode.
It may not be possible to perform the control due to the limit of the control range and the time required for setting.

A method of controlling the magnitude of the image output signal by changing the light quantity on the irradiation light source side and changing the light quantity on the image pickup surface of the TV camera as the image pickup means has also been considered. That is, a method of controlling the light emission amount of the light source and a method of inserting an optical filter or a diaphragm for changing the light amount into a part of the optical system are adopted.

Furthermore, since the image pickup area changes depending on the measurement mode, it is necessary to increase or decrease the width of the sample flow itself. Further, since the image magnification is changed every time the measurement mode is changed, it is necessary to switch the objective lens or the projection lens.

If the number of measurement modes is about two, the optical system can be configured including the switching mechanism section. However, if there are more measurement modes than these, these optical systems are configured.・ It is difficult to place. In addition to the time required for switching the optical system, the focusing operation required for switching the optical system is also required.

Further, in the low magnification measurement mode, the peripheral light amount of the picked-up image is insufficient, so-called shading occurs in the image signal, which causes obstacles in the image processing stage. Usually, a means for improving the flatness is provided in the light irradiation system, and shading correction is electrically performed.

Further, in the case of optically dealing with it, a light amount control mechanism is necessary because the light amount becomes insufficient in the high magnification measurement mode. On the other hand, the electrical shading correction requires a complicated work of creating data for the shading correction and performing a correction calculation based on the data for each image.

As described above, switching the imaging magnification in each measurement mode causes an increase in processing time and device cost, affects the configuration of each part of the device, and complicates the configuration of the particle image analyzer. And there is a problem of making it expensive.

The present invention is based on the above-mentioned prior art dead time,
It was made to solve various problems caused by switching of multiple measurement modes, and there is no dead time.
Solves the complexity of the optical system, mechanical system, and image processing system configuration that occurs when switching between measurement modes in multiple measurement modes, and provides multiple types of accurate and high-precision regardless of particle concentration. It is an object of the present invention to provide a flow-type particle image analysis device capable of determining the particle classification and the number of particles.

[0026]

In order to achieve the above object, the structure of the flow type particle image analysis apparatus according to the present invention is as follows:
Flow a particle sample suspended in a liquid into a flow cell,
A particle image analyzing apparatus for morphologically classifying particles, comprising: an image capturing section that captures particles that have passed through an image capturing area in the flow cell; and an image analyzing section that analyzes an image of a particle image by the image capturing section. Then, the imaging unit has a plurality of measurement modes, continuously captures the particle images by the same imaging system in different measurement modes, and the image analysis unit captures the continuous particle images. The analysis is performed for each pixel column.

When the particle image is continuous in a plurality of pixel rows, the image analysis section analyzes each pixel row and adds the analysis value for each pixel row. The imaging unit is configured to change the method of imaging particles. The imaging method is changed by changing either one or both of the imaging region and the flow path width of the flow cell. The image pickup unit has the same image pickup magnification of the image pickup system. The imaging unit is configured to perform particle detection and particle classification by one device. The imaging unit is a one-dimensional sensor.

The particle sample is characterized in that it is a biological cell. The particle sample is a blood cell component present in blood. The particle sample is characterized in that it is urine, particularly a urine sediment component present in urine.

[0029]

The function of each of the above technical means is as follows.
According to the configuration of the present invention, a separate optical system for particle detection is not provided, and an optical system having a simple configuration allows an imaging region in a particle sample suspended in a liquid flowing in a flow cell. The passed particles can be continuously imaged, and the imaged particle image can be subjected to image analysis for each pixel line, and the morphological classification of the particles can be performed.

Further, since all the images of the sample liquid containing the particles flowing through the flow cell are imaged and the image processing of the particles is performed, the particles can be counted without counting, and the particles can be counted regardless of the concentration of the particles. And the particle concentration can be accurately calculated. In a plurality of measurement modes, it is not necessary to change the optical system of the image pickup unit, and thus it is not necessary to readjust the optical system of the image pickup unit.

In the image taken for each line of pixels, the presence of particles can be determined by image processing, and at the same time, the characteristics of the particles can be extracted if the particles are present.
If no particles are detected in one pixel line, the image processing is ended as it is, and the image processing of the next pixel line is started.
If particles are present in one pixel line, the feature amount per one pixel line is calculated.

When the particle image is continuous in a plurality of pixel rows, the feature amount of the particle portion of the first pixel line is calculated,
The feature amount of the next pixel line is calculated and added to the feature amount of the first pixel line. If a particle image also exists in the next pixel line, the feature amount of particles is added up to that line. As long as particles are present in one pixel line, such processing can be continued, and the result can be output when the particles are not present in the pixel line.

If such image processing for each line is performed in real time, feature extraction can be performed for all particles regardless of particle concentration. Further, the number of particles and the particle concentration can be accurately obtained.

When the measurement mode is changed and the image magnification is changed, the image pickup area becomes smaller in the case of high magnification than in the case of low magnification. Therefore, if switching is performed such that imaging is performed by a partial area of the pixel sensor when the magnification is high and imaging is performed by the entire area of the pixel sensor when the magnification is low, the optical system can be quickly changed without changing the optical system. It can be imaged. The width of the flow path of the flow cell is also changed along with the change of the imaging region. In addition, when the magnification is low, the all-pixel sensor is not used, and the pixel sensor is used at a predetermined interval so that the image pickup and the image processing can be performed quickly.

[0035]

Embodiments of the present invention will be described below with reference to FIGS. 1 to 3. [Embodiment 1] FIG. 1 is a block diagram showing the overall configuration of a particle image analysis apparatus according to an embodiment of the present invention, and FIG.
FIG. 3 is a block diagram of a particle feature extraction unit in the particle image analysis apparatus according to the embodiment of FIG. 3, and FIG. 3 is a principal block diagram when the embodiment of FIG. 1 is applied to slide glass sample analysis.

In FIG. 1, 1 is a light source for imaging a sample image, 4 is a field lens, 5 is a flow cell through which a sample solution flows, 6 is a sample solution in a flow cell 5, and 7 is a sample solution. Objective lens, 8 is a slit, 9 is a one-dimensional CC for imaging
D and 14 are AD converters, 20 is an image processing unit, 27
Is a memory, 29 is a flow control unit, 31 is a display, 32 is a printer, and 35 is a control computer.

In this embodiment, the sample solution 6 in which particles are suspended is flown through the flow cell 5, the light source 1 is constantly turned on and the flow cell 5 is irradiated, and the flow direction of the flow cell 5 is changed. An image of the sample liquid 6 passing through the image capturing area in the flow cell 5 is continuously captured by a one-dimensional sensor 9 provided on a plane parallel to the flow direction and is obtained. The image is processed for each pixel line, and the particles in the sample liquid 6 are classified.

A halogen lamp is actually used as the imaging light source 1. For the sake of explanation, it is assumed that the sample liquid flows perpendicularly to the paper surface from the front side to the back side.
The sample liquid 6 flows through the flow cell 5 extremely flatly, and its flow rate is kept constant by the flow control unit 29.

The flow path of the flow cell 5 is the halogen lamp 1
Has a dimension of about 100 to 200 μm in the direction parallel to the irradiation optical axis (horizontal direction in FIG. 1) and has a dimension of several mm in the direction perpendicular to the irradiation optical axis (vertical direction in FIG. 1). . The light emitted from the halogen lamp 1 is collimated by the lens 4 and is applied to the flow cell 5. The light that has passed through the flow cell 5 is condensed by the objective lens 7, forms an image on the slit 8, and then passes therethrough.

The light passing through the slit 8 is converted into an electric signal by the one-dimensional CCD 9 for image pickup. The one-dimensional C
The CD 9 is arranged at a right angle to the sample liquid 6 on a plane parallel to the flow of the sample liquid 6.

Since the one-dimensional CCD 9 constantly monitors the sample liquid 6 and repeatedly images it, the one-dimensional CC 9
A voltage corresponding to the amount of light accumulated for each pixel is sequentially output from D9. In this way, the image of the sample liquid 6 is continuously output for each one-line pixel column. 1
The output signal for each line pixel column is converted to A by the AD converter 14.
D-converted and input to the image processing unit 20.

FIG. 2 is an example of a block diagram of an image processing unit in the particle image analysis apparatus according to the embodiment of FIG. In FIG. 2, the image processing unit 20 includes a binary image processing unit 150.
A grayscale image processing unit 151 and a pixel coordinate processing unit 152
And a feature extraction unit 172 and a particle determination unit 174, the output of which is sent to the memory 27.

Further, the feature extraction unit 172 can carry out the perimeter length calculation processing 160, the projection length calculation processing 161, the area calculation processing 162, the maximum / minimum coordinate calculation processing 163, and the density calculation processing 164. ing. A digital signal from the AD converter 14 is input, and the signal becomes a binary image signal in the binary image processing unit 150.

Perimeter calculation processing 16 from the binary image signal
0, projection length calculation processing 161, and area calculation processing 162
And maximum and minimum coordinate calculation processing 163 and density calculation processing 16
4 is applied and the parameters of perimeter, projected length, particle area, projected length, maximum and minimum coordinates, and density are calculated. Also, A
The digital signal from the D converter 14 is the grayscale image processing unit 1
The density calculation processing 51 is performed, and the density calculation processing 16 is performed based on the density signal.
4 is done and the density of the particle image is calculated. Similarly, the pixel coordinate processing unit 152 and the maximum / minimum coordinate calculation processing 16
The maximum and minimum coordinates are calculated by 3 and.

Each feature amount calculated by the feature extraction unit 172 is input in parallel to the particle determination unit 174, and the particle determination unit 174 determines the measurement particle type from each input feature amount. , Memory 27 is output. The memory 27 reads the analysis result stored in real time into the control computer 35 when the measurement of the sample is completed, and outputs the analysis result to the display 32 and the printer 31.

These characteristic amount calculation processes are carried out in parallel by known one-pass characteristic extraction. Further, since the feature amount calculation and the particle determination are performed in real time, they are configured by so-called pipeline processing. Therefore, since the feature amount calculation for one pixel line is executed by one clock pulse, the feature extraction unit 72 can be configured by hardware logic.

First, when the particle size is large and the image covers a plurality of pixel columns, first. The feature amount of the particle portion of the first pixel line is calculated. Next, the feature amount of the pixel line following the first pixel line is calculated, and the feature amounts of both pixel lines are added. If a particle image also exists in the next pixel line, the particle feature amount of that line is calculated and further added to the added feature amount of both pixel lines. As long as there are particles in one pixel line, such processing is continued. The process is terminated when the particles are no longer present in the pixel line, and is output to the memory 27. In this way, the feature amount can be quickly calculated even for large-sized particles.

The measurement mode is changed by changing the one-dimensional CCD 9
This is performed by changing the image pickup area and the flow channel width of the flow cell 5. Although not shown, the flow channel width is changed by inserting a weir plate into the flow channel and moving the weir plate by a drive unit.

Next, the known one-pass feature extraction will be described. Details of 1-pass feature extraction are described in "Yoda, Inai, Ejiri," Video Rate 1-pass Method for Particle Image Analysis ", 1988 Autumn Meeting of the Institute of Electronics, Information and Communication Engineers. A brief description will be given by giving it to a known document.

In the one-pass feature extraction, when the image information is continuously input once from the image pickup device line by line, the feature amount of the image to be measured is calculated in real time and has the following features. . (1) An input image is input line by line according to clock pulses at regular intervals. (2) All the processing required when one line is input can be executed with one clock pulse. That is, even if multiple processes are required, they are performed in parallel with one clock pulse. (3) The individual data to be output is output at the time when the operation of the image area relating to it is completed, or after a fixed clock from that time, whereby the processing concerning the output data is completed.

As a method of calculating the feature amount, a label is attached to a graphic element portion intersecting each line, and the label is propagated to the intersecting intersection on the next line while The partial feature amount is accumulated on the label.

The present invention is not limited to this example, and can be applied to the case where the sample is particles in a slide glass sample. This will be described with reference to FIG. Figure 3
FIG. 4 is a block diagram of essential parts when applied to analysis of a slide glass sample of the present invention.

In FIG. 3, 100 is a control circuit and 102 is a control circuit.
Is a slide glass sample, and 104 is a measurement sample. A measurement sample 104 is obtained by moving the slide glass sample 102 at a constant speed in the direction of the arrow shown by the control circuit 100.
Particles can be subjected to particle detection and image analysis in the same manner as particles in the flow cell.

The embodiment thus constructed has the following effects. 1. Since the optical system can be configured with one optical system, the arrangement of the optical systems can be compactly integrated. 2. The transmission information of the sample flow can be acquired by the one-dimensional line sensor, and the image processing can be performed in real time. Therefore, detection omission of particles can be prevented, and the measurement accuracy is significantly improved. 3. In a plurality of measurement modes, the same optical system is used when changing the measurement mode, so that there is no need to readjust the optical system. 4. Since the one-dimensional line sensor images all the transmission information of the sample flow containing particles for each line, no particle omission occurs. 6. Since the one-dimensional line sensor is used, the cost can be reduced as compared with the one using the TV camera.

[0055]

As described in detail above, according to the present invention, there is no dead time, and the optical system, the mechanical system, and the image processing system, which are caused by switching the measurement modes in a plurality of measurement modes, are generated. It is possible to provide a flow-type particle image analyzer capable of solving the complexity of each configuration and accurately and accurately determining a plurality of types of particle classifications and the number of particles regardless of the particle concentration.

[Brief description of drawings]

FIG. 1 is a block diagram showing the overall configuration of a particle image analysis apparatus that is an embodiment of the present invention.

FIG. 2 is a block diagram of a particle feature extraction unit in the particle image analysis apparatus according to the embodiment of FIG.

FIG. 3 is a block diagram of a main part when the embodiment of FIG. 1 is applied to an analysis of a slide glass sample.

[Explanation of symbols]

 1 Halogen lamp 4 Field lens 5 Flow cell 6 Sample 7 Objective lens 8 Slit 9 One-dimensional CCD 14 AD converter 20 Image processing unit 27 Memory 29 Flow control unit 31 Display 32 Printer 35 Control computer 100 Slide glass control unit 102 Slide glass 104 Sample 150 Binary image processing unit 151 Gray image processing unit 152 Pixel coordinate processing unit 160 Perimeter length processing 161 Projection length processing 162 Area processing 163 Maximum and minimum coordinate processing 164 Density processing 172 Feature extraction unit 174 Particle determination unit

Claims (11)

[Claims] 1. A particle sample suspended in a liquid is flowed.
An image capturing unit that captures the particles that have flowed through the cell and that has passed through the image capturing region in the flow cell, and an image analyzing unit that analyzes the image of the particle image by the image capturing unit are provided. In the -type particle image analysis device, the image capturing unit has a plurality of measurement modes, continuously captures the particle images by the same image capturing system in different measurement modes, and the image analyzing unit A flow characterized in that it is configured to analyze the continuous particle image for each pixel column.
Particle image analyzer 2. The image analysis unit is configured to analyze each particle row when a particle image is continuous in a plurality of pixel rows and add the analysis value for each pixel row. The flow type particle image analysis device according to claim 1. 3. The flow-type particle image analysis device according to claim 1, wherein the image pickup unit is configured to change a method of picking up particles. 4. The flow-type particle image analysis according to claim 3, wherein the imaging method is changed by changing either one of the imaging area and the flow path width of the flow cell. apparatus. 5. The flow-type particle image analysis apparatus according to claim 3, wherein the change of the imaging method is configured to change the imaging region and the flow channel width of the flow cell. 6. The flow-type particle image analysis apparatus according to claim 1, wherein the image pickup section is configured so that the image pickup systems have the same image pickup magnification. 7. The flow chart according to claim 1, wherein the image pickup unit is configured to perform image pickup for particle detection and image pickup for particle classification by one image pickup unit. Particle image analyzer. 8. The flow-type particle image analysis device according to claim 1, wherein the imaging unit is composed of a one-dimensional sensor. 9. The flow-type particle image analyzer according to claim 1, wherein the particle sample is a biological cell. 10. The flow-type particle image analysis device according to claim 1, wherein the particle sample is a blood cell component present in blood. 11. The flow-type particle image analyzer according to claim 1, wherein the particle sample is urine, particularly a urinary sediment component present in urine.