CN100353758C - Coding method for monitoring image resolution and its apparatus - Google Patents

Abstract

The present invention relates to an encoding method and device for monitoring image resolution. The steps of the encoding method are as follows: 1) take odd and even digital video images, and convert them into a frame of D1 image by deinterlacing; 2) perform border and zoom processing on the image , converted into a frame of PAL format: 528 pixels x 384 pixels or NTSC format: 528 pixels x 320 pixels. 3) Use the H.264 compression algorithm to compress the image in step 2) into a frame of DCIF image. The device includes a camera, a video conversion unit for converting the analog video signal input by the camera into a digital video signal, a memory, a video encoding and information processing unit. The subjective quality of the image encoded and output by the present invention is good, and in the case of an output code stream of 384Kbps to 1024Kbps, it better satisfies the requirements of users in the digital monitoring industry that the video playback quality is good and the details are clear.

Description

Coding method and device for monitoring image resolution

technical field

The invention relates to a coding method and a device for monitoring image resolution, and belongs to the field of data image compression processing.

Background technique

As we all know, to measure the quality of a digital surveillance image, in addition to the subjective feeling, it also includes the following important technical indicators: the resolution of the output image, the signal-to-noise ratio of the output image, and the size of the data stream of the output image.

At present, the following resolutions are mainly used in the video surveillance industry:

QCIF (quarter common intermediate format, quarter common intermediate format) (PAL, Phase Alternating Line, progressive phase inversion, 176 pixels × 144 pixels; NationalTelevision System Committee, National Television System Committee, 176 pixels *120 pixels);

CIF (common intermediate format, common intermediate format) (PAL: 352 pixels × 288 pixels; NTSC: 352 pixels × 240 pixels);

HALF D1 (half first digital video format, one-half of the first digital video format) (PAL: 720 pixels × 288 pixels; NTSC: 720 pixels × 240 pixels);

D1 (first digital video format, the first digital video format) (PAL: 720 pixels × 576 pixels; NTSC: 720 pixels × 480 pixels), these resolutions have deficiencies.

1) QCIF resolution Although the encoded data stream is small, but the image pixels are few, the image resolution is not high, and many details are often unclear, so it is only used in a small amount of network monitoring occasions where the transmission bandwidth is strictly limited.

2) CIF resolution is currently the mainstream application resolution, and the data flow is moderate (384Kbps to 1024Kbps). If the signal-to-noise ratio is above 32dB, it is acceptable for general monitoring users, but the video image quality is still not ideal, and the image resolution is still Not high enough.

3) Compared with CIF resolution, HALF D1 resolution only increases the pixels in the horizontal direction, and does not change in the vertical direction. The improvement of image quality is not particularly obvious, but the code stream increases greatly, and it is only used in some products.

4) D1 resolution image has good clarity and meets the high quality requirements, but at the cost of high bit rate and high price. With the current encoding algorithm, when its bit rate is reduced to less than 1024Kbps, the video image quality is not as good as CIF resolution, and the encoding cost increases significantly. It is only used in some high-end occasions.

Digital surveillance applications usually require long-term, multi-channel video recording. Due to the limitation of storage space and transmission bandwidth, the video code stream should not be higher than the CIF resolution coded output code stream, that is, 384Kbps to 1024Kbps, higher than this code flow, unbearable. Therefore, how to improve the video playback image quality while maintaining the 384Kbps to 1024Kbps output stream is one of the urgent problems to be solved in the current digital surveillance industry, especially in important surveillance sites.

Contents of the invention

The purpose of the present invention is to provide a coding method and device for improving the image resolution of the monitoring video image quality under the condition of 384Kbps to 1024Kbps output code stream.

The coding method of monitoring image resolution of the present invention comprises the following steps:

PAL format:

1) the odd and even two fields in the digital video signal, each field of 720 pixels × 288 pixels, are subjected to anti-interlacing conversion to form an image of a frame of 720 pixels × 576 pixels;

2) The image obtained in step 1) is subjected to boundary processing, and the leftmost and rightmost 8 pixels of each row are removed to become an image of a frame of 704 pixels × 576 pixels;

3) The image obtained in step 2) is scaled, reduced by 3/4 horizontally, and reduced by 2/3 vertically, and converted into an image of a frame of 528 pixels * 384 pixels;

4) For the image obtained in step 3), use the H.264 compression algorithm to compress it into an H.264 image frame of 528 pixels×384 pixels.

NTSC format:

1) the odd and even two fields in the digital video signal, each field of 720 pixels × 240 pixels, are subjected to anti-interlacing conversion to form a frame of images of 720 pixels × 480 pixels;

2) The image obtained in step 1) is subjected to boundary processing, and the leftmost and rightmost 8 pixels of each row are removed to become an image of a frame of 704 pixels * 480 pixels;

3) The image obtained in step 2) is scaled, reduced by 3/4 horizontally, and reduced by 2/3 vertically, and converted into an image of 528 pixels * 320 pixels in a frame;

4) For the image obtained in step 3), use the H.264 compression algorithm to compress it into a video frame of 528 pixels×320 pixels.

Taking the PAL system as an example, because the total number of pixels of the image resolution 528 * 384 of the present invention is just in time twice of the total number of pixels of CIF (352 * 288), in order to be different from the commonly said 2CIF (704 pixels * 288 pixels) Point, also called Half D1) resolution distinction, below we will be called DCIF (DOUBLE CIF, double CIF) with the image resolution of the present invention's encoding.

The device for realizing the above monitoring image resolution encoding method includes:

A camera for capturing analog video;

The video conversion unit is used to convert the analog video signal output by the camera into a digital video signal and output it to the video encoding and information processing unit. The input end of the video conversion unit is connected to the camera, and the output end of the video conversion unit is connected to the video encoding and information processing unit. unit connected;

The memory is connected to the video encoding and information processing unit, and stores the temporary data of the video encoding and information processing unit;

The video encoding and information processing unit processes and compresses the odd and even fields of video image data taken from the memory into one frame of DCIF resolution video frame.

Advantages of the invention

(1) Subjectively feel that the image quality is good. In the security industry that generally uses CIF as the compression resolution, the DCIF encoding of the present invention basically exchanges the bandwidth of CIF resolution for better than CIF resolution, which is basically equivalent to the image quality of 4CIF resolution, without giving storage and network Transmissions bring extra stress.

(2) Compared with the Half D1 resolution, the DCIF (528×384) coding resolution of the present invention is more balanced in both horizontal and vertical directions.

(3) The resolution of the DCIF encoding of the present invention can well take into account the requirements of users in the digital monitoring industry for higher image quality and lower code stream: that is, the playback image is required to have high definition and meet the requirements for details; the amount of data is required to be small , to ensure that the system can record for a long time and stabilize real-time network transmission.

Description of drawings

Fig. 1 is the encoding device of monitoring image resolution of the present invention;

Fig. 2 is a flowchart of the implementation steps of the encoding method for monitoring image resolution of the present invention;

Figure 3 is a schematic diagram of the experimental results of the relationship between the weighted peak signal-to-noise ratio and the output code stream;

Table 1 is the experimental result of the relationship between the weighted peak signal-to-noise ratio and the output code stream corresponding to Fig. 3 .

Detailed ways

With reference to Fig. 1, coding device of the present invention comprises the analog video signal that camera 1 inputs is converted into the video conversion unit 2 of digital video signal, memory 3 and video coding and information processing unit 4, the input terminal of video conversion unit 2 and camera 1 The output terminal of the video converting unit 2 and the memory 3 are respectively connected to the video encoding and information processing unit 4 . The video encoding and information processing unit 4 may use a PNX1500 series media processor chip from Philips of the Netherlands. The video conversion unit 2 may use a SAA7113H chip or a SAA7144HL chip.

The video encoding and information processing unit 4 of the device is connected with the host computer/main CPU unit 6 connected with the hard disk 5 and the network interface 7 through the PCI bus. During work, the video conversion unit 2 converts the analog image data from the camera 1 into a digital signal suitable for processing by the video encoding and information processing unit 4, and stores it in the memory 3; the video encoding and information processing unit 4 gets parity from the memory 3 Two video (PAL system: 720 pixels × 288 pixels or NTSC system: 720 pixels × 240 pixels) image data and process the data series, use H.264 compression algorithm to compress it into a DCIF resolution format (PAL system: 528 pixels × 384 pixels or NTSC system: 528 pixels × 320 pixels) video frame, temporarily stored in the memory; stored in the memory The video frame sequence is sent to the host/main CPU unit 6 through the PCI bus under the control of the video encoding and information processing unit 4, and is stored on the hard disk 5 by the host or the main CPU unit 6 or transmitted to the remote end through the network interface 7.

The encoding method of the monitoring image resolution of the present invention is implemented by the video encoding and information processing unit 4 in the device through a program, and the program flow is as shown in Figure 2:

1) read the image data of odd and even fields (under the PAL system: 720 pixels × 288 pixels or under the NTSC system: 720 pixels × 240 pixels) image data from the memory, and judge whether it is DCIF encoding, If yes, go to step 2; if not, perform other resolution encoding process.

2) to the odd and even two field image data that step 1) obtains, through deinterlacing (deinterlace) conversion, form a frame D1 (under the PAL system: 720 pixel points * 576 pixel points; under the NTSC system: 720 pixel points dot × 480 pixel dot) image;

3) Do boundary processing on the D1 image generated in step 2), that is, for each row, remove the leftmost and rightmost 8 pixels, and become a frame of 4CIF (PAL system: 704 pixels × 576 pixels dot; under NTSC system: 704 pixels × 480 pixels) image;

4) The 4CIF image produced in step 3) is scaled, that is, the horizontal 3/4 is reduced, and the vertical 2/3 is reduced, and converted into a frame (under the PAL system: 528 pixel points × 384 pixel points; under the NTSC system: 528 pixel × 320 pixel) image.

5) a frame (under the PAL system: 528 pixel points * 384 pixel points; under the NTSC system: 528 pixel points * 320 pixel points) image generated in step 4), use the H.264 compression algorithm to compress it Compressed into one frame under PAL system: 528 pixels × 384 pixels or under NTSC system: 528 pixels × 320 pixels.

Repeat steps 1-5 to realize the output of continuous DCIF resolution video code stream.

Figure 3 is a schematic diagram of the experimental results of the relationship between WPSNR (weighted peak signal-to-noise ratio, weighted peak signal-to-noise ratio) and the output code stream, and Table 1 is the experimental results of the relationship between WPSNR and the output code stream corresponding to Figure 3. As can be seen from the figure, under the 384kbps output code stream situation, the WPSNR of the DCIF code of the present invention is 1dB higher than the WPSNR of the CIF code, and 2.5dB higher than the WPSNR of the 4CIF code; The WPSNR of WPSNR is 2.2dB higher than that of 4CIF encoding, which is 0.5dB higher than that of 4CIF encoding; in the case of 768kbps output stream, the WPSNR of DCIF encoding is nearly 3dB higher than that of CIF encoding, which is basically the same as that of 4CIF encoding; in the case of 1024kbps output stream The WPSNR of DCIF encoding is 3.28dB higher than that of CIF encoding, and 0.5dB lower than that of 4CIF encoding.

Table 1

Note: Since 4CIF cannot be compressed to 192K and 256K, their weighted peak signal-to-noise ratio is not calculated

Claims (4)

1. the coding method of a monitoring image resolution is characterized in that may further comprise the steps:

Pal mode:

1) with two of strange, the idols in the digital video signal, every 720 picture element * 288 picture elements carry out anti-interlacing conversion, form the image of a frame 720 picture elements * 576 picture elements;

2) image that step 1) is obtained is done BORDER PROCESSING, removes the most left and the rightest each 8 picture element of each row, becomes the image of a frame 704 picture elements * 576 picture elements;

3) to step 2) image that obtains carries out convergent-divergent and handles, and level 3/4 is dwindled, and vertical 2/3 dwindles, and converts the image of a frame 528 picture elements * 384 picture elements to;

4) image that step 3) is obtained uses H.264 compression algorithm, and it is compressed into the H.264 picture frame of a frame 528 picture element x 384 picture elements.

The TSC-system formula:

1) with two of strange, the idols in the digital video signal, every 720 picture element * 240 picture elements carry out anti-interlacing conversion, form the image of a frame 720 picture elements * 480 picture elements;

2) image that step 1) is obtained is done BORDER PROCESSING, removes the most left and the rightest each 8 picture element of each row, becomes the image of a frame 704 picture elements * 480 picture elements;

3) to step 2) image that obtains carries out convergent-divergent and handles, and level 3/4 is dwindled, and vertical 2/3 dwindles, and converts the image of a frame 528 picture elements * 320 picture elements to;

4) image that step 3) is obtained uses H.264 compression algorithm, and it is compressed into the frame of video of a frame 528 picture elements * 320 picture elements.

2. enforcement of rights requires the device of 1 described monitoring image resolution coding method, it is characterized in that comprising:

Camera (1) is used to gather analog video;

Video converting unit (2), be used for converting the analog video signal of camera (1) output to digital video signal, be defeated by video coding and information process unit (4), the input of video converting unit (2) links to each other with camera (1), and the output of video converting unit (2) links to each other with video coding and information process unit (4);

Memory (3) links to each other with video coding and information process unit (4), the temporal data of store video coding and information process unit (4);

Video coding and information process unit (4) will be taken from two vedio datas of odd even of memory (3) and handle the frame of video that is compressed into a frame DCIF resolution format.

3. device according to claim 2 is characterized in that video coding and information process unit (4) are PNX1500 series media processor chips.

4. device according to claim 2 is characterized in that video converting unit (2) is SAA7113H chip or SAA7144HL chip.

Images