CN106791843A - A kind of Lossless Image Compression Algorithm system and method - Google Patents

Abstract

The invention provides a lossless image compression system, comprising a state register unit, a data distributor, an S-O-F inserter, a difference encoding unit and a code stream synthesizer, and a state register unit for registering and outputting at least Each serial number of a frame of images; a data distributor for receiving said images, and gating an S-O-F inserter or a difference coding unit according to said serial number; an S-O-F inserter and a difference coding unit , output the processed data to the code stream synthesizer, and the code stream synthesizer integrates and sequentially selects and outputs the compressed image information according to the serial number. The invention also provides a lossless image compression method, which effectively reduces the spatial redundancy in the image, facilitates image alignment and fast retrieval, and has a simple algorithm with a high degree of parallelism, and is easy to transplant to the hardware of the algorithm. Real-time image compression without loss of image information.

Description

An image lossless compression system and method

technical field

The invention relates to the technical field of image information processing, in particular to an image lossless compression system and method.

Background technique

Image sensors have been widely used in various fields such as scientific research, industrial production, medical and health care, national defense and military affairs. With the continuous improvement of image resolution and image frame rate, human beings can observe more subtle object features or more short-term physical phenomena, which greatly enriches human observation methods and improves human cognition.

However, with the rapid improvement of image performance parameters, it also brings a huge challenge that the amount of image data and data transmission bandwidth increase exponentially. Massive image data and ultra-high data transmission bandwidth have brought a series of technical difficulties for subsequent image transmission, storage and processing.

At present, it is relatively difficult to perform real-time compression without information loss for such high-resolution, ultra-high-bandwidth image data streams at high frame rates (especially for image data streams exceeding 10 Gbps). Limited by high frame rate, high data bandwidth, and processing capability, most of the existing high-speed image acquisition systems store raw image data directly without any processing. Therefore, there is a large amount of redundant data in the image data, which actually wastes storage capacity, occupies additional transmission bandwidth, and greatly increases the implementation cost of the imaging system.

Contents of the invention

(1) Technical problems to be solved

The object of the present invention is to provide an image lossless compression system and method to solve at least one of the above technical problems.

(2) Technical solutions

The invention provides an image lossless compression system, comprising:

A status register unit, configured to register and output each sequence number of at least one frame of images received;

a data distributor, configured to receive the image, and select an S-O-F inserter or a difference encoding unit according to the serial number;

S-O-F inserter (data frame header inserter), used for outputting the dark pixel row and the effective pixel row first row of the first frame image according to the sequence number, and the effective pixels of the remaining frames except the first frame image row the first row; and replace the dark pixel rows of each frame image except the first frame image with the alignment code, and output the alignment code;

The difference-seeking coding unit is used to calculate the residuals of the remaining effective pixel rows before and after the remaining effective pixel rows of each frame of image except the first effective pixel row; encode the residual, and output the encoded marked value and overflow value;

The code stream synthesizer is used for integrating and sequentially gating and outputting the compressed image information according to the sequence number.

Preferably, the compressed image information is the marked value and overflow value of the dark pixel row of the first frame image, the first row of effective pixel row, and the rest of the effective pixel row; The alignment code of the first row of effective pixel rows, the dark pixel row, and the label value and overflow value of the remaining effective pixel rows.

Preferably, the state register unit includes:

The loop counter is used to register the pixel serial number of the image data;

The loop counter is used to store the row number of the image data;

Incremental counter, used to register the frame number of image data.

Preferably, the difference-seeking coding unit includes:

a data double buffering unit for receiving, storing and outputting the image;

An inter-frame difference calculator, used to obtain and output the residual and overflow value;

A difference encoder, configured to encode the residual, and output an encoded marked value;

The overflow value data buffer unit is used for storing and outputting the overflow value.

Based on the same inventive concept, the present invention also provides a method for image lossless compression, including:

S1. Receive at least one frame of image in sequence, register and output each serial number of the image, that is, pixel serial number, row serial number and frame serial number;

S2. Processing the first frame of image: output the dark pixel row of the first frame image and the first row of effective pixels; calculate and encode the remaining effective pixel rows to calculate the residuals of each row, obtain and output the coded marked value and overflow value;

S3. Process the remaining frames of images except the first frame of images: obtain and output the alignment codes of the dark pixel rows of the remaining frames of images; output the first row of effective pixels; calculate and encode the residuals of the remaining effective pixel rows before and after the calculation, Obtain and output encoded marked value and overflow value;

S4. According to the sequence number, integrate and sequentially select and output the compressed image information.

Preferably, the compressed image information is the marked value and overflow value of the dark pixel row of the first frame image, the first row of effective pixel row, and the rest of the effective pixel row; The alignment code of the dark pixel row, the first row of the valid pixel row, and the label value and overflow value of the remaining valid pixel rows.

Preferably, the alignment code satisfies the formula:

or

Among them, gray(p, l, f) is the grayscale function of the image, p is the pixel number of the image, 1 is the row number, f is the frame number, M is the image width, N is the image height, and R is the quantization of the image pixel Bit width, l _dark is the dark pixel row height of the image, % is the remainder operator.

Preferably, the residual δ(p, l, f) satisfies the formula:

δ(p,l,f)=gray(p,l,f)-gray(p,l-1,f)

(1≤p≤M, L _dark +1<l≤N, f≥1).

Preferably, the indicated value B(p, l, f) satisfies the formula:

Among them, TH is the threshold value, TH=2 ^k -1 (0<K<R); binK _{δ(p, l, f)} is the K-bit signed natural binary code of the residual, and binK _TH is the K-bit signed natural binary code of TH Natural binary code, binK _-(TH+1) is a K-bit signed natural binary code of -(TH+1).

Preferably, the overflow value O(p, l, f) satisfies the formula:

Among them, binR _{δ(p, l, m)} is the R-bit signed natural binary code of the residual.

(3) Beneficial effects

1. The image lossless compression method of the present invention is simple and efficient, the hardware implementation time delay is low, and a higher processing frequency can be realized; in addition, the parallelism is high, and multiple frames of images and multiple image blocks can be processed, so it is very beneficial The hardware transplantation of the algorithm can realize real-time compression under extremely high data bandwidth;

2. The image lossless compression system of the present invention includes an S-O-F inserter, which can not only retain dark pixel data, but also set the frame header alignment code of the image, which is beneficial to rapid image retrieval, positioning and alignment in the post-processing of compressed data or image reconstruction stage;

3. The present invention is applicable to the compression of high frame rate image data streams and slowly changing image data streams;

4. The algorithm of the present invention can realize real-time image compression under extremely high data bandwidth without loss of image information.

Description of drawings

Fig. 1 is the structural representation of the embodiment of the present invention;

Fig. 2 is a schematic diagram of steps of an embodiment of the present invention;

Fig. 3 is the specific implementation flowchart of the embodiment of the present invention;

4 is a schematic diagram of an output data stream format according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of a test image of an embodiment of the present invention.

detailed description

An aspect of the embodiments of the present invention provides an image lossless compression system, including:

A status register unit, configured to register and output each sequence number of at least one frame of images received;

Specifically, the state registering unit includes: a loop counter, used for registering the pixel serial number of the image data; a loop counter, used for registering the row serial number of the image data; an increment counter, used for registering the frame serial number of the image data.

a data distributor, configured to receive the image, and select an S-O-F inserter or a difference encoding unit according to the serial number;

The S-O-F inserter is used to output the first row of dark pixels and the first row of effective pixel rows of the first frame image, and the first row of effective pixel rows of each frame image except the first frame image according to the sequence number; and Replace the dark pixel rows of each frame image except the first frame image with the alignment code, and output the alignment code;

The difference-seeking coding unit is used to calculate the residuals of the remaining effective pixel rows before and after the remaining effective pixel rows of each frame of image except the first effective pixel row; encode the residual, and output the encoded marked value and overflow value;

Specifically, the difference encoding unit includes: a data double buffer unit for receiving, storing and outputting the image; an inter-frame difference calculator for calculating and outputting the residual and overflow value; the difference An encoder, configured to encode the residual, and output an encoded flag value; an overflow value data buffer unit, configured to store and output the overflow value.

A code stream synthesizer, used to integrate and sequentially strobe and output compressed image information according to the sequence number: the dark pixel row of the first frame image, the first row of valid pixel rows, and the marked values and overflow of the remaining valid pixel rows Value; the alignment codes of the first line of effective pixel rows and dark pixel rows of the other frame images except the first frame image, and the label value and overflow value of the remaining effective pixel rows.

Another aspect of the embodiments of the present invention also provides a method for image lossless compression, including:

S1. Receive at least one frame of image in sequence, register and output each serial number of the image, that is, pixel serial number, row serial number and frame serial number;

S2. Processing the first frame of image: output the dark pixel row of the first frame image and the first row of effective pixels; calculate and encode the remaining effective pixel rows to calculate the residuals of each row, obtain and output the coded marked value and overflow value;

S3. Process the other frames of images except the first frame of images: obtain and output the alignment codes of the dark pixel rows of the remaining frames of images;

Output the first row of effective pixels; calculate and encode the residuals of the remaining effective pixel rows before and after the calculation, and obtain and output the coded marked value and overflow value;

The alignment code satisfies the formula:

or

Among them, gray(p, l, f) is the grayscale function of the image, p is the pixel number of the image, l is the line number, f is the frame number, M is the image width, N is the image height, and R is the quantization of the image pixel Bit width, l _dark is the dark pixel row height of the image, % is the remainder operator.

S4. According to the serial number, integrate and sequentially strobe and output the compressed image information: the dark pixel row of the first frame image, the first row of effective pixel rows, and the marked values and overflow values of the remaining effective pixel rows; except for the first Alignment codes of the first row of effective pixel rows and dark pixel rows of the remaining frames of images outside the frame image, and the label values and overflow values of the remaining effective pixel rows.

Wherein, the residual δ (p, l, f) satisfies the formula:

δ(p,l,f)=gray(p,l,f)-gray(p,l-1,f)

(1≤p≤M, L _dark +1<l≤N, f≥1).

The indicated value B(p, l, f) satisfies the formula:

Among them, TH is the threshold value, TH=2 ^k -1 (0<K<R); binK _{δ(p, l, f)} is the K-bit signed natural binary code of the residual, and binK _TH is the K-bit signed natural binary code of TH Natural binary code, binK _-(TH+1) is a K-bit signed natural binary code of -(TH+1).

The overflow value O(p, l, f) satisfies the formula:

Among them, binR _{δ(p, l, m)} is the R-bit signed natural binary code of the residual.

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

Fig. 1 is a schematic structural diagram of the implementation of the present invention. As shown in Fig. 1, the image lossless compression system includes: a state register unit, a data distributor, an S-O-F inserter, a difference encoding unit and a code stream synthesizer. Described state register comprises P loop counter, is used for registering the pixel serial number p of image data; L loop counter, is used for registering row serial number 1 of image data; With F incremental counter, is used for registering the frame serial number f of image data; The unit receives at least one frame of image, registers and outputs the serial numbers of the image, namely pixel serial number p, row serial number 1 and frame serial number f, to the data distributor, S-O-F inserter and code stream synthesizer.

The data distributor receives the image and the p, l, f sequence numbers output by the status register unit to select the S-O-F inserter or the difference encoding unit. The S-O-F inserter receives the p, l, f serial numbers, outputs the first row of dark pixel rows and the first row of effective pixel rows of the first frame image, and the first row of effective pixel rows of all other frame images except the first frame image; And replace the dark pixel rows of the other frame images except the first frame image with the alignment code, and output the alignment code G(p, l, f). The difference encoding unit includes: a data double buffer unit, an inter-frame difference calculator, a difference encoder and an overflow value data buffer unit. The data double buffering unit receives, stores and outputs the image; the inter-frame difference operator calculates and outputs the residual δ (p, l, f) and overflow value O (p, l, f); difference The encoder encodes the residual, and outputs the encoded label value B(p, l, f); the overflow value data buffer unit stores and outputs the overflow value to the code stream synthesizer.

The code stream synthesizer receives the information output by the S-O-F inserter, the difference encoder and the overflow value data buffer, and outputs the compressed image information according to the sequence number: the dark pixel row of the first frame image, the first row of the effective pixel row , and the marked value and overflow value of the other valid pixel rows; the alignment code of the dark pixel row of the other frame images except the first frame image, the first row of the effective pixel row, and the marked value and overflow value of the remaining valid pixel rows .

Fig. 2 is a schematic diagram of the steps of an embodiment of the present invention, as shown in Fig. 2, the image lossless compression method includes:

S1. Receive at least one frame of image in sequence, register and output each serial number of the image, that is, pixel serial number, row serial number and frame serial number;

S2. Processing the first frame of image: output the dark pixel row of the first frame image and the first row of effective pixels; calculate and encode the remaining effective pixel rows to calculate the residuals of each row, obtain and output the coded marked value and overflow value;

The residual satisfies the formula:

δ(p,l,f)=gray(p,l,f)-gray(p,l-1,f)

(1≤p≤M, L _dark +1<l≤N, f≥1)

The indicated value B(p, l, f) satisfies the formula:

Among them, TH is the threshold value, TH=2 ^k -1 (0<K<R); binK _{δ(p, l, f)} is the K-bit signed natural binary code of the residual, and binK _TH is the K-bit signed natural binary code of TH Natural binary code, binK _-(TH+1) is a K-bit signed natural binary code of -(TH+1).

The overflow value O(p, l, f) satisfies the formula:

Among them, binR _{δ(p, l, m)} is the R-bit signed natural binary code of the residual.

S3. Process the remaining frames of images except the first frame of images: obtain and output the alignment codes of the dark pixel rows of the remaining frames of images; output the first row of effective pixels; calculate and encode the residuals of the remaining effective pixel rows before and after the calculation, Obtain and output encoded marked value and overflow value;

The alignment code satisfies the formula:

or

Among them, gray(p, l, f) is the grayscale function of the image, p is the pixel number of the image, l is the line number, f is the frame number, M is the image width, N is the image height, and R is the quantization of the image pixel Bit width, l _dark is the dark pixel row height of the image, % is the remainder operator.

S4. According to the serial number, integrate and sequentially strobe and output the compressed image information: the dark pixel row of the first frame image, the first row of effective pixel rows, and the marked values and overflow values of the remaining effective pixel rows; except for the first The alignment codes of the dark pixel rows, the first row of effective pixel rows, and the indication values and overflow values of the remaining effective pixel rows of the remaining frames of images outside the frame image.

Fig. 3 is the specific implementation flowchart of the embodiment of the present invention, as shown in Fig. 3, the image lossless compression system starts to process at least one frame image, first judges whether the current image is the first frame image according to the frame number f registered in the increment counter, Then compare the row number 1 stored in the loop counter L with the dark pixel row l _dark to determine whether the current row is a dark pixel row or each row of a valid pixel row. When the current image is the first frame image: if the current line is the first line of the dark pixel line or the effective pixel line, then directly output the original data to the code stream synthesizer; if the current line is valid except for the first line of effective pixel line For a pixel row, calculate the residual of this row and the previous row, then judge the residual and the threshold TH, and output the corresponding marked value and overflow value. If the current image is a frame image other than the first frame: if the current row is a dark pixel row, calculate and output the alignment code; if the current row is the first row of a valid pixel row, output the original data directly without processing; if the current row For the remaining effective pixel rows except the first effective pixel row, calculate the residual of this row and the previous row, then judge the residual and threshold TH, and output the corresponding marked value and overflow value to the code stream synthesizer. Finally, the code stream synthesizer integrates the received data and outputs compressed image information.

Fig. 4 is a schematic diagram of the output data stream format of the embodiment of the present invention. As shown in Fig. 4, the compressed information output by the image lossless compression system is: the dark pixel row of the first frame image, the first row of effective pixel row , and the marked value and overflow value of the other valid pixel rows; the alignment code of the dark pixel row of the other frame images except the first frame image, the first row of the effective pixel row, and the marked value and overflow value of the remaining valid pixel rows .

Fig. 5 is the test image schematic diagram of the embodiment of the present invention, as shown in Fig. 5, original image is 2560 (H) * 2162 (V) * 16bits=88555520bits, dark pixel line 2 lines, through compression method processing of the present invention, set The fixed threshold is 127, that is, K=8, the theoretical limit compression ratio is 1.997, the actual compressed image size is 44883689bits, and the image compression ratio is about 1.973.

The specific embodiments described above have further described the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above descriptions are only specific embodiments of the present invention, and are not intended to limit the present invention. Within the spirit and principles of the present invention, any modifications, equivalent replacements, improvements, etc., shall be included in the protection scope of the present invention.

Claims (10)

1. A lossless image compression system, comprising: A status register unit, configured to register and output each sequence number of at least one frame of images received; a data distributor, configured to receive the image, and select an S-O-F inserter or a difference encoding unit according to the serial number; The S-O-F inserter is used to output the first row of dark pixels and the first row of effective pixel rows of the first frame image, and the first row of effective pixel rows of each frame image except the first frame image according to the sequence number; and Replace the dark pixel rows of each frame image except the first frame image with the alignment code, and output the alignment code; The difference-seeking coding unit is used to calculate the residuals of the remaining effective pixel rows before and after the remaining effective pixel rows of each frame of image except the first effective pixel row; encode the residual, and output the encoded marked value and overflow value; The code stream synthesizer is used for integrating and sequentially gating and outputting the compressed image information according to the sequence number. 2. The system according to claim 1, wherein the compressed image information is the marked value and overflow value of the dark pixel row of the first frame image, the first row of effective pixel rows, and the remaining effective pixel rows ; Alignment codes of the first line of effective pixel rows and dark pixel rows of the other frames of images except the first frame of image, and the marked values and overflow values of the remaining effective pixel rows. 3. The system according to claim 1, wherein the state register unit comprises: The loop counter is used to register the pixel serial number of the image data; The loop counter is used to store the row number of the image data; Incremental counter, used to register the frame number of image data. 4. The system according to claim 1, wherein the difference encoding unit comprises: a data double buffering unit for receiving, storing and outputting the image; An inter-frame difference calculator, used to obtain and output the residual and overflow value; A difference encoder, configured to encode the residual, and output an encoded marked value; The overflow value data buffer unit is used for storing and outputting the overflow value. 5. A method for image lossless compression, comprising: S1. Receive at least one frame of image in sequence, register and output each serial number of the image, that is, pixel serial number, row serial number and frame serial number; S2. Processing the first frame of image: output the dark pixel row of the first frame image and the first row of effective pixels; calculate and encode the remaining effective pixel rows to calculate the residuals of each row, obtain and output the coded marked value and overflow value; S3. Process the remaining frames of images except the first frame of images: obtain and output the alignment codes of the dark pixel rows of the remaining frames of images; output the first row of effective pixels; calculate and encode the residuals of the remaining effective pixel rows before and after the calculation, Obtain and output encoded marked value and overflow value; S4. According to the sequence number, integrate and sequentially select and output the compressed image information. 6. The method according to claim 5, wherein the compressed image information is the marked value and overflow value of the dark pixel row of the first frame image, the first row of effective pixel rows, and the remaining effective pixel rows ; The alignment codes of the dark pixel rows, the first row of valid pixel rows, and the marked values and overflow values of the remaining valid pixel rows except the first frame of image. 7. The method according to claim 5, wherein the alignment code satisfies the formula: or Among them, gray(p, l, f) is the grayscale function of the image, p is the pixel number of the image, l is the line number, f is the frame number, M is the image width, N is the image height, and R is the quantization of the image pixel Bit width, l _dark is the dark pixel row height of the image, % is the remainder operator. 8. method according to claim 7, is characterized in that, described residual δ (p, l, f) satisfies formula: δ(p,l,f)=gray(p,l,f)-gray(p,l-1,f) (1≤p≤M, L _dark +1<l≤N, f≥1). 9. method according to claim 8, is characterized in that, described mark value B (p, l, f) satisfies formula: Among them, TH is the threshold value, TH=2 ^k -1 (0<K<R); binK _{δ(p, l, f)} is the K-bit signed natural binary code of the residual, and binK _TH is the K-bit signed natural binary code of TH Natural binary code, binK _-(TH+1) is a K-bit signed natural binary code of -(TH+1). 10. method according to claim 9, is characterized in that, described overflow value O (p, l, f) satisfies formula: Among them, binR _{δ(p, l, m)} is the R-bit signed natural binary code of the residual.