Multi-dimensional optimized audio and video lossless compression method, system, computer device and storage medium

Abstract

The application discloses a kind of multi-dimension optimized audio and video lossless compression method, system, computer equipment and storage medium.The method comprises: audio and video sequence is collected;The audio and video sequence is divided into key frame and non-key frame;For the key frame, multi-dimension information optimization is carried out;The key frame is compressed;Non-key frame data is optimized based on the multi-dimension optimization information obtained in the third step;The non-key frame data after optimization is compressed.The application realizes the improvement of compression efficiency while ensuring the high fidelity of audio and video content.The method not only carries out fine processing for key frame and non-key frame respectively, but also optimizes through brightness and chroma block, effectively reduces data redundancy by combining high-density hybrid prediction.This innovative compression technology can be widely applied to various audio and video processing scenarios, such as streaming media transmission, video storage and real-time communication, etc.

Description

Technical Field

[0001] The present invention belongs to the field of audio and video compression technology, and in particular to a multi-dimensional optimized audio and video lossless compression method, system, computer equipment and storage medium. Background Art

[0002] Audio and video data compression technology is a key research area in digital media processing. With the rapid development of the internet, the generation and dissemination of audio and video content has exploded. To effectively store and transmit this massive amount of data, developing efficient compression algorithms is crucial. Compression technology must not only minimize data redundancy while maintaining video quality, but also support real-time transmission and playback. In recent years, the application of emerging technologies such as deep learning has provided new insights into audio and video compression, driving advancements and developments in related technologies.

[0003] Traditional audio and video compression technologies are primarily based on methods such as discrete cosine transform (DCT), motion compensation, and predictive coding. Common coding standards such as H.264 and H.265 analyze the similarities between video frames and exploit the differences between keyframes (I-frames) and non-keyframes (P-frames and B-frames) to reduce redundant information. In practice, traditional methods first divide the video frame into blocks, then perform a DCT transform on each block to extract frequency domain information and quantize it. While these technologies reduce data volume to a certain extent, they still face a trade-off between compression ratio and image quality when processing complex scenes.

[0004] While traditional compression technology has achieved some success in video data processing, it still suffers from several significant drawbacks. First, traditional methods often struggle to maintain high audio and video fidelity in highly dynamic scenes or complex image content, resulting in a degradation of image quality after compression. Second, existing technologies handle redundant data in a crude manner, failing to fully utilize the multi-dimensional characteristics of images, resulting in an ineffective reduction of data redundancy. Furthermore, traditional methods place high demands on computing resources, particularly in real-time applications, which can lead to delays and insufficient processing speeds. Summary of the Invention

[0005] The purpose of the present invention is to provide an audio and video lossless compression method based on multi-dimensional optimization to address the problems existing in the above-mentioned prior art.

[0006] The technical solution to achieve the purpose of the present invention is: a multi-dimensional optimized audio and video lossless compression method, the method comprising:

[0007] Step 1: Collect audio and video sequences;

[0008] Step 2: dividing the audio and video sequence into key frames and non-key frames;

[0009] Step 3: Optimize multi-dimensional information for the key frame;

[0010] Step 4, compressing the key frame;

[0011] Step 5: Optimize non-keyframe data based on the multi-dimensional optimization information obtained in step 3;

[0012] Step 6: compress the optimized non-key frame data.

[0013] Furthermore, in step 3, multi-dimensional information optimization is performed on the key frame, specifically including:

[0014] Step 3-1, convert the key frame image into RGB-HSL-YUV;

[0015] Step 3-2: Divide the image converted in step 3-1 into multiple pixel blocks, and characterize the brightness and chromaticity information of each pixel block to obtain a brightness prediction value and a chromaticity prediction value of each pixel block;

[0016] Step 3-3, performing DCT transformation on the image processed in step 3-2.

[0017] Furthermore, the conversion process in step 3-1 further includes:

[0018] Setting a first quantization threshold;

[0019] If any component of the pixel is less than the first quantization threshold, it is set to 0.

[0020] Furthermore, in step 3-2, for each pixel block, a prediction model is constructed based on the surrounding pixel information, and the luminance prediction value and the chrominance prediction value are obtained using the prediction model; the prediction model is specifically:

[0021] ;

[0022] Where, Represents the brightness / chrominance prediction value of pixel block (i, j), N represents the total number of pixels in pixel block (i, j), Represents the brightness value of the kth pixel in the pixel block, is the corresponding weight, and is adaptively adjusted according to the local gradient information of the pixel.

[0023] Furthermore, step 5 specifically includes:

[0024] Step 5-1, convert the non-keyframe image into RGB-HSL-YUV to extract the brightness information, chroma information and the chroma-to-brightness residual;

[0025] Step 5-2: Divide the image converted in step 5-1 into multiple pixel blocks, and extract the pixel mean in the horizontal and vertical directions for each pixel block;

[0026] Step 5-3: In the luminance domain, for each pixel block, calculate the error between the actual pixel value and the luminance prediction value and chrominance prediction value of the key frame. The calculation formula is:

[0027]

[0028] Where E represents the error value, I represents the actual pixel value, that is, the pixel mean extracted in step 5-2, and P represents the brightness prediction value or chrominance prediction value of the key frame;

[0029] Step 5-4: Use the deep model to reconstruct the brightness information, chrominance information, and chrominance-to-luminance residual of non-key frames. The specific process includes:

[0030] The multi-dimensional optimization information of the key frame obtained in step 2 and the error value obtained in step 5-3 are used to train the deep learning model; the multi-dimensional optimization information of the key frame is used for supervised learning to construct a nonlinear mapping relationship between brightness information, chrominance information, and chrominance to brightness residual;

[0031] The luminance information, chrominance information, and chrominance-to-luminance residual of non-key frames are input into the trained deep learning model to achieve reconstruction;

[0032] Step 5-5, performing DCT transformation on the image reconstructed in step 5-4.

[0033] Furthermore, in step 5-1, the calculation formula for the chroma to luminance residual is:

[0034]

[0035] Where R represents the residual from chroma to luminance, C represents the extracted chroma information, and Y represents the luminance information of the key frame after optimization in step 3. A function that represents the influence of brightness information on chromaticity, using linear or nonlinear mapping.

[0036] Furthermore, steps 4 and 6 compress the key frames according to the video standard to ensure that the compressed data conforms to the standard format.

[0037] Furthermore, in step 4 or step 6, the data after DCT transformation is compressed using an entropy coding method based on a video standard. The specific process includes:

[0038] Step 4-1: quantize the DCT transform coefficient matrix. The quantization process controls the accuracy of different frequency components through the quantization matrix. The formula is as follows:

[0039]

[0040] in, Represents the quantized DCT transform coefficients, is the original DCT transform coefficient, is the corresponding element of the standard quantization matrix;

[0041] Step 4-2, arranging the quantized DCT transform coefficients in a "Z-shaped" scanning order to form a one-dimensional data stream;

[0042] Step 4-3, use Huffman coding to encode the quantized DCT transform coefficients, construct a variable-length coding table based on the probability of data occurrence, assign relatively shorter codes in the variable-length coding table to high-frequency data, and assign relatively longer codes in the variable-length coding table to low-frequency data.

[0043] In another aspect, a multi-dimensionally optimized audio and video lossless compression system is provided, the system comprising:

[0044] The first module is used to collect audio and video sequences;

[0045] The second module is used to divide the audio and video sequence into key frames and non-key frames;

[0046] The third module is used to perform multi-dimensional information optimization on the key frame;

[0047] A fourth module is used to compress the key frame;

[0048] A fifth module is used to optimize non-keyframe data based on the multi-dimensional optimization information obtained in the third module;

[0049] The sixth module is used to compress the optimized non-key frame data.

[0050] On the other hand, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the multi-dimensionally optimized audio and video lossless compression method when executing the computer program.

[0051] Compared with the prior art, the present invention has the following significant advantages:

[0052] (1) Improving compression efficiency: By adopting multi-dimensional information optimization and block processing technology, the generation of redundant data is significantly reduced. During keyframe processing, the conversion from RGB to HSL and HSL to YUV is used to optimize the extraction of brightness and chrominance information, minimizing data redundancy. Non-keyframes further optimize data representation by extracting residual and depth information. This multi-level compression strategy significantly reduces the size of the final compressed file, greatly saving storage space and bandwidth, and improving data transmission efficiency.

[0053] (2) Maintaining high fidelity: During the compression process, luminance and chrominance block optimization technology is used to ensure the restoration of image details and colors. By setting a reasonable quantization threshold, the pixel values ​​in flat areas are effectively set to zero, while areas with rich details retain more information, ensuring the clarity and realism of the image. In this way, even under high compression ratios, users can still experience visual effects close to the original audio and video content, meeting the high fidelity requirements.

[0054] (3) Flexibility and Adaptability: By introducing a deep learning model to extract and reconstruct image features, the proposed solution possesses strong flexibility and adaptability. During training, the model can adaptively learn the features of different content and better capture image details when processing complex scenes. This makes the solution applicable not only to static images but also to dynamic videos, making it widely adaptable to a wide range of different types of audio and video data.

[0055] (4) Reduced computational complexity: This solution effectively simplifies the computational requirements during the compression process by combining block optimization with high-density hybrid prediction. The discrete cosine transform (DCT) and its optimized form are used to convert the time domain signal into a frequency domain representation, effectively suppressing high-frequency noise. During secondary prediction, the previous frame information can be used for efficient prediction, thereby reducing the calculation of redundant data and improving overall processing speed, making it suitable for real-time application scenarios.

[0056] (5) Broad Application Prospects: The innovative and efficient solution of the present invention has broad application prospects, especially in the fields of streaming media transmission, video storage, and real-time communication. In streaming media platforms, users' requirements for video quality and smoothness are constantly increasing, and the lossless compression technology provided by this solution can meet this demand, ensuring no distortion or loss during the transmission process. In addition, in high-definition video storage, the high compression ratio of this solution will help users save storage costs and achieve more efficient content management.

[0057] The present invention is further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A schematic diagram of the principles of a multi-dimensionally optimized audio and video lossless compression method in one embodiment.

[0059] Figure 2 FIG. 1 is a schematic diagram of performing multi-dimensional information optimization on key frames in one embodiment.

[0060] Figure 3 FIG. 1 is a schematic diagram of optimizing non-key frame data based on multi-dimensional optimization information in one embodiment. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0062] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back, etc.), such directional indications are only used to explain the relative position relationship, movement status, etc. between the various components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0063] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features specified as "first" and "second" may explicitly or implicitly include at least one of such features. In addition, the technical solutions between the various embodiments can be combined with each other, but this must be based on the fact that ordinary technicians in this field can implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0064] In one embodiment, combined Figure 1 , provides a multi-dimensional optimized audio and video lossless compression method, the method comprising:

[0065] Step 1: Collect audio and video sequences;

[0066] Step 2: dividing the audio and video sequence into key frames and non-key frames;

[0067] Step 3: Optimize multi-dimensional information for the key frame;

[0068] Step 4, compressing the key frame;

[0069] Step 5: Optimize non-keyframe data based on the multi-dimensional optimization information obtained in step 3;

[0070] Here, non-key frame data is optimized based on the multi-dimensional optimization information of the key frames to improve the efficiency of subsequent compression and the visual quality after decoding.

[0071] Step 6: compress the optimized non-key frame data.

[0072] Furthermore, in one embodiment, in combination Figure 2 In step 3, multi-dimensional information optimization is performed on the key frame, specifically including:

[0073] Step 3-1, convert the key frame image into RGB-HSL-YUV;

[0074] Here, the image is converted from RGB space to HSL and YUV color space. This conversion helps to separate brightness and color information and improve the processing ability of data of different dimensions.

[0075] Preferably, the conversion process further includes: setting a first quantization threshold; and if any component of a pixel is less than the first quantization threshold, setting it to 0. This ensures that low-intensity noise does not affect subsequent processing.

[0076] Here, the first quantization threshold is set based on the statistical distribution of the image. The optimized color representation reduces unnecessary information redundancy and mitigates the impact of high-frequency noise. After conversion, the image data's luminance information (Y) and chrominance information (U, V) in YUV space are optimized separately. Luminance information (Y) is used for subsequent structural analysis, while chrominance information (U, V) is used for further residual information extraction.

[0077] Step 3-2: Divide the image converted in step 3-1 into multiple pixel blocks, and characterize the brightness and chromaticity information of each pixel block to obtain a brightness prediction value and a chromaticity prediction value of each pixel block;

[0078] Here, the pixel blocks are divided based on the adaptive block size, which can maximize the preservation of local features while ensuring computational efficiency.

[0079] Step 3-3, performing DCT transformation on the image processed in step 3-2.

[0080] Here, DCT transform can convert pixel data from the time domain to the frequency domain, so that the energy of the image is more concentrated in the low-frequency components, thereby improving the data compression rate. The formula of DCT transform is as follows:

[0081] .

[0082] in, Represents the frequency domain coefficient after DCT transformation, is the pixel value in the spatial domain, MMM and N represent the number of rows and columns of the image block respectively, and is a normalization factor that ensures consistent energy in the transformation. The core goal of the DCT transform is to concentrate energy in the low-frequency portion, making subsequent encoding and compression more efficient while reducing non-essential information and improving the overall compression ratio. Combined with the aforementioned luminance and chrominance optimization, this step ensures that the keyframe's visually significant features are preserved as much as possible during the compression process, laying the foundation for subsequent keyframe compression.

[0083] Preferably, in some embodiments, in step 3-2, for each pixel block, a prediction model is constructed based on the surrounding pixel information, and the prediction model is used to obtain the brightness prediction value and the chrominance prediction value to reduce data redundancy; the prediction model is specifically:

[0084]

[0085] Where, Represents the brightness / chrominance prediction value of pixel block (i, j), N represents the total number of pixels in pixel block (i, j), Represents the brightness value of the kth pixel in the pixel block, is the corresponding weight, and is adaptively adjusted according to the local gradient information of the pixel.

[0086] The solution of this embodiment can effectively reduce the brightness error of key frames, improve prediction accuracy, and provide better basic data for subsequent DCT transformation.

[0087] Furthermore, in one embodiment, in combination Figure 3 , step 5 specifically includes:

[0088] Step 5-1, convert the non-keyframe image into RGB-HSL-YUV to extract the brightness information, chroma information and the chroma-to-brightness residual;

[0089] Here, by calculating the chroma-to-luminance residual, the expression of chroma information is further refined and the accuracy of color restoration is improved.

[0090] Here, the calculation formula for the chroma to luma residual is:

[0091]

[0092] Where R represents the residual from chroma to luminance, C represents the extracted chroma information, and Y represents the luminance information of the key frame after optimization in step 3. A function representing the effect of luminance on chrominance, typically using linear or nonlinear mapping, makes the chrominance residual information more representative. This residual information is then fed into the deep learning model as supplementary information during the subsequent reconstruction process, enhancing the color reproduction of non-keyframes.

[0093] Step 5-2: Divide the image converted in step 5-1 into multiple pixel blocks, and extract the pixel mean in the horizontal and vertical directions for each pixel block to enhance the perception of local patterns in the image;

[0094] Step 5-3: In the luminance domain, for each pixel block, calculate the error between the actual pixel value and the luminance prediction value and chrominance prediction value of the key frame. The calculation formula is:

[0095]

[0096] Where E represents the error value, I represents the actual pixel value, that is, the pixel mean extracted in step 5-2, and P represents the brightness prediction value or chrominance prediction value of the key frame;

[0097] Step 5-4: Use the deep model to reconstruct the brightness information, chrominance information, and chrominance-to-luminance residual of non-key frames. The specific process includes:

[0098] The multi-dimensional optimization information of the key frame obtained in step 2 and the error value obtained in step 5-3 are used to train the deep learning model; the multi-dimensional optimization information of the key frame is used for supervised learning to construct a nonlinear mapping relationship between brightness information, chrominance information, and chrominance to brightness residual;

[0099] The luminance information, chrominance information, and chrominance-to-luminance residual of non-key frames are input into the trained deep learning model to achieve reconstruction;

[0100] Here, the error value is used to train the deep learning model, enabling it to more accurately predict the brightness and chrominance information of non-key frames, reduce reconstruction errors, and improve video quality.

[0101] Here, reconstruction is performed based on the learned key frame features to make the data quality of non-key frames as close to the original frames as possible.

[0102] Step 5-5, performing DCT transformation on the image reconstructed in step 5-4.

[0103] Here, the spatial domain information is converted into frequency domain information through DCT transformation for subsequent compression encoding, ensuring that the data remains efficient and conforms to the standard format during storage and transmission.

[0104] Furthermore, in one embodiment, steps 4 and 6 compress the key frames according to the video standard to ensure that the compressed data conforms to the standard format. At the same time, the multi-dimensional information optimized in step 3 is combined during the compression process to improve compression efficiency and reconstruction quality.

[0105] Preferably, in some embodiments, after DCT transformation, the key frame data has been converted into frequency domain representation, wherein the low-frequency components contain the main image information and the high-frequency components store the details and edge information.

[0106] Therefore, in step 4 or step 6, the DCT-transformed data is compressed using an entropy coding method based on video standards. This ensures that the compressed data conforms to the standard format while minimizing data redundancy and improving storage and transmission efficiency. Common entropy coding methods include run-length coding and variable-length coding. Run-length coding reduces data storage by recording the number of consecutive identical values, while variable-length coding allocates codes of varying lengths based on the probability of occurrence to reduce the storage overhead of high-frequency data.

[0107] The specific process of the entropy coding method includes:

[0108] Step 4-1: quantize the DCT transform coefficient matrix to reduce the data accuracy of high-frequency components and improve the compression rate. The quantization process controls the accuracy of different frequency components through the quantization matrix. The formula is as follows:

[0109]

[0110] in, Represents the quantized DCT transform coefficients, is the original DCT transform coefficient, is the corresponding element of the standard quantization matrix;

[0111] Here, the quantization matrix is ​​optimized based on the visual characteristics of the human eye. A smaller quantization step size is used for low-frequency components to retain more image details, while a larger quantization step size is used for high-frequency components to reduce the amount of data storage while ensuring that the compressed image still has a high perceptual quality.

[0112] Step 4-2: Arrange the quantized DCT transform coefficients in a zigzag scanning order to form a one-dimensional data stream for subsequent entropy coding;

[0113] Step 4-3: Encode the quantized DCT transform coefficients using Huffman coding, construct a variable-length coding table based on the probability of data occurrence, assign relatively shorter codes in the variable-length coding table to high-frequency data, and assign relatively longer codes in the variable-length coding table to low-frequency data;

[0114] Here, the Huffman coding process is based on statistical information, calculating the probability distribution of different DCT coefficients and constructing a binary tree based on this information. The coding rules follow the no-prefix principle to ensure unique data decoding. The resulting encoded data meets video standards, ensuring correct parsing at the video decoder. Combined with the aforementioned optimization process, keyframes retain sufficient visual information after compression, providing a foundation for subsequent optimization of non-keyframes.

[0115] In one embodiment, a multi-dimensional optimized audio and video lossless compression system is provided, the system comprising:

[0116] The first module is used to collect audio and video sequences;

[0117] The second module is used to divide the audio and video sequence into key frames and non-key frames;

[0118] The third module is used to perform multi-dimensional information optimization on the key frame;

[0119] A fourth module is used to compress the key frame;

[0120] A fifth module is used to optimize non-keyframe data based on the multi-dimensional optimization information obtained in the third module;

[0121] The sixth module is used to compress the optimized non-key frame data.

[0122] For the specific definition of the multi-dimensional optimized audio and video lossless compression system, please refer to the definition of the multi-dimensional optimized audio and video lossless compression method above, which will not be repeated here. The various modules in the above-mentioned multi-dimensional optimized audio and video lossless compression system can be implemented in whole or in part by software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0123] In one embodiment, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following is achieved:

[0124] Step 1: Collect audio and video sequences;

[0125] Step 2: dividing the audio and video sequence into key frames and non-key frames;

[0126] Step 3: Optimize multi-dimensional information for the key frame;

[0127] Step 4, compressing the key frame;

[0128] Step 5: Optimize non-keyframe data based on the multi-dimensional optimization information obtained in step 3;

[0129] Step 6: compress the optimized non-key frame data.

[0130] For the specific limitations of each step, please refer to the limitations of the multi-dimensional optimized audio and video lossless compression method above, which will not be repeated here.

[0131] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the computer program implements:

[0132] Step 1: Collect audio and video sequences;

[0133] Step 2: dividing the audio and video sequence into key frames and non-key frames;

[0134] Step 3: Optimize multi-dimensional information for the key frame;

[0135] Step 4, compressing the key frame;

[0136] Step 5: Optimize non-keyframe data based on the multi-dimensional optimization information obtained in step 3;

[0137] Step 6: compress the optimized non-key frame data.

[0138] For the specific limitations of each step, please refer to the limitations of the multi-dimensional optimized audio and video lossless compression method above, which will not be repeated here.

[0139] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only illustrative of the principles of the present invention. Without departing from the spirit and scope of the present invention, any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included in the scope of protection of the present invention.

Claims

1. A multi-dimensional optimized audio and video lossless compression method, characterized in that: The method comprises: Step 1: Collect audio and video sequences; Step 2: dividing the audio and video sequence into key frames and non-key frames; Step 3: Optimize multi-dimensional information for the key frame; Step 4, compressing the key frame; Step 5: Optimize non-keyframe data based on the multi-dimensional optimization information obtained in step 3; Step 6: compress the optimized non-key frame data; Step 5 specifically includes: Step 5-1, convert the non-keyframe image into RGB-HSL-YUV to extract the brightness information, chroma information and the chroma-to-brightness residual; Step 5-2: Divide the image converted in step 5-1 into multiple pixel blocks, and extract the pixel mean in the horizontal and vertical directions for each pixel block; Step 5-3: In the luminance domain, for each pixel block, calculate the error between the actual pixel value and the luminance prediction value and chrominance prediction value of the key frame. The calculation formula is: ； Where E represents the error value, I represents the actual pixel value, that is, the pixel mean extracted in step 5-2, and P represents the brightness prediction value or chrominance prediction value of the key frame; Step 5-4: Use the deep model to reconstruct the brightness information, chrominance information, and chrominance-to-luminance residual of non-key frames. The specific process includes: The multi-dimensional optimization information of the key frame obtained in step 3 and the error value obtained in step 5-3 are used to train the deep learning model; the multi-dimensional optimization information of the key frame is used for supervised learning to construct a nonlinear mapping relationship between brightness information, chrominance information, and chrominance to brightness residual; The luminance information, chrominance information, and chrominance-to-luminance residual of non-key frames are input into the trained deep learning model to achieve reconstruction; Step 5-5, performing DCT transformation on the image reconstructed in step 5-4; In step 5-1, the calculation formula for the chroma to luminance residual is: ； Where R represents the residual from chroma to luminance, C represents the extracted chroma information, and Y represents the luminance information of the key frame after optimization in step 3. A function that represents the influence of brightness information on chromaticity, using linear or nonlinear mapping.

2. The multi-dimensional optimized audio and video lossless compression method according to claim 1, characterized in that: In step 3, multi-dimensional information optimization is performed on the key frame, specifically including: Step 3-1, convert the key frame image into RGB-HSL-YUV; Step 3-2: Divide the image converted in step 3-1 into multiple pixel blocks, and characterize the brightness and chromaticity information of each pixel block to obtain a brightness prediction value and a chromaticity prediction value of each pixel block; Step 3-3, performing DCT transformation on the image processed in step 3-2.

3. The multi-dimensional optimized audio and video lossless compression method according to claim 2, characterized in that: The conversion process in step 3-1 also includes: Setting a first quantization threshold; If any component of the pixel is less than the first quantization threshold, it is set to 0.

4. The multi-dimensional optimized audio and video lossless compression method according to claim 2, characterized in that: In step 3-2, for each pixel block, a prediction model is constructed based on the surrounding pixel information, and the luminance prediction value and the chrominance prediction value are obtained using the prediction model; the prediction model is specifically: ； Where, Represents the brightness / chrominance prediction value of pixel block (i, j), N represents the total number of pixels in pixel block (i, j), Represents the brightness value of the kth pixel in the pixel block, is the corresponding weight, and is adaptively adjusted according to the local gradient information of the pixel.

5. The multi-dimensional optimized audio and video lossless compression method according to claim 1, characterized in that: Steps 4 and 6 compress the key frames according to the video standard to ensure that the compressed data conforms to the standard format.

6. The multi-dimensional optimized audio and video lossless compression method according to claim 2, characterized in that: In step 4 or step 6, the data after DCT transformation is compressed using an entropy coding method based on a video standard. The specific process includes: Step 4-1: quantize the DCT transform coefficient matrix. The quantization process controls the accuracy of different frequency components through the quantization matrix. The formula is as follows: ； in, Represents the quantized DCT transform coefficients, is the original DCT transform coefficient, is the corresponding element of the standard quantization matrix; Step 4-2, arranging the quantized DCT transform coefficients in a "Z-shaped" scanning order to form a one-dimensional data stream; Step 4-3, use Huffman coding to encode the quantized DCT transform coefficients, construct a variable-length coding table based on the probability of data occurrence, assign relatively shorter codes in the variable-length coding table to high-frequency data, and assign relatively longer codes in the variable-length coding table to low-frequency data.

7. A multi-dimensional optimized audio and video lossless compression system based on the method according to any one of claims 1 to 6, characterized in that: The system comprises: The first module is used to collect audio and video sequences; The second module is used to divide the audio and video sequence into key frames and non-key frames; The third module is used to perform multi-dimensional information optimization on the key frame; A fourth module is used to compress the key frame; A fifth module is used to optimize non-keyframe data based on the multi-dimensional optimization information obtained in the third module; The sixth module is used to compress the optimized non-key frame data.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.

Images