Digital signal real-time processing system based on multi-core distributed processing

Abstract

The invention discloses a digital signal real-time processing system based on multi-core distributed processing, which relates to the technical field of digital signal processing and comprises a data receiving module, a parallel processing module, a process scheduling module and a terminal interaction module. Identifying a coding protocol of the input video digital signal, receiving the input video digital signal through an external interface of a preset multi-communication protocol, and processing the input video digital signal to obtain a to-be-analyzed video frame sequence; performing color space conversion processing on a to-be-analyzed video frame sequence to obtain a grayscale image sequence, a hue component and a saturation component, performing calculation through two-dimensional fast Fourier transform and a cascade filter bank to obtain enhanced video frame grayscale image data, and performing pseudo-color mapping to obtain an enhanced video frame image; and according to the current frame image and the previous frame image of the to-be-analyzed video frame sequence, calculating to obtain a current frame motion feature map and a moving target positioning mask, so that digital signal processing is efficiently completed at the front end to obtain an enhanced image and target mask positioning.

Description

Technical Field

[0001] The present invention relates to the technical field of digital signal processing, and in particular to a digital signal real-time processing system based on multi-core distributed processing. Background Art

[0002] In recent years, information technology has developed rapidly, and the performance of hardware such as dedicated central processing units (CPUs), digital signal processors (DSPs), and graphics processing units (GPUs) has achieved great progress, providing efficient computing support for the rapid implementation of complex algorithms. At the same time, computing methods such as multi-core distributed processing have also improved the real-time processing efficiency of video image digital signals.

[0003] At present, the Chinese invention patent application with application number CN2021105265278 discloses an image signal processor and an image sensor processing system. The application includes: in the processor, the reconfigurable interface module can be configured through compilation, read the original image data from the outside of the image signal processor, and send the original image data to the reconfigurable digital signal processing module. The hardware architecture or hardware connection inside the reconfigurable digital signal processing module can be restructured through software compilation to meet the hardware implementation requirements of the image processing algorithm, process the original image data, and have edge computing capabilities. Therefore, the image signal processor provided by the application can be adapted to image sensors of different types or specifications, has edge computing capabilities, and meets the needs of algorithm processing based on artificial intelligence theoretical models. However, the application lacks targeted enhancement processing of video image data, and the direct visual effect of the video image obtained by the post-receiving terminal is not good, which increases the processing time of the post-receiving terminal. Summary of the invention

[0004] The technical problem solved by the present invention is that the traditional method lacks targeted enhancement processing of video image data, the direct visual effect of the video image obtained by the post-receiving terminal is not good, and the processing time of the post-receiving terminal is increased.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: A digital signal real-time processing system based on multi-core distributed processing, comprising: Data receiving module, parallel processing module, process scheduling module and terminal interaction module; The data receiving module includes a signal input unit and a signal buffer unit, which are used to identify the encoding protocol of the input video digital signal, receive the input video digital signal through an external interface of a preset multi-communication protocol, and process the input video digital signal through a first buffer block, a frame extraction intermediate block and a second buffer block to obtain a video frame sequence to be analyzed; The parallel processing module includes a color space conversion unit, a filtering enhancement unit and a target extraction unit, which are used to perform color space conversion processing on a video frame sequence to be analyzed to obtain a grayscale image sequence, hue H component data and saturation S component data corresponding to the video frame sequence to be analyzed, perform two-dimensional fast Fourier transform and double-layer filtering processing of a cascade filter group on the video frame grayscale image data to calculate enhanced video frame grayscale image data, perform pseudo-color mapping on the enhanced video frame grayscale image data to obtain an enhanced video frame image, calculate a current frame motion feature map according to a current frame image and a previous frame image of the video frame sequence to be analyzed, and obtain a moving target positioning mask through binarization processing and morphological processing; The process scheduling module is used to monitor the resource utilization and fault conditions of the processors in the signal receiving module, the filter enhancement unit and the moving target identification unit, and to allocate and schedule the processing processes according to the preset processing priority rules and user instructions; The terminal interaction module is used to receive the user's operation instruction data and set the system preset parameters, and upload the enhanced video frame image and the moving target positioning mask corresponding to each video frame in the video frame sequence to be analyzed to the cloud server.

[0006] As a preferred solution of the digital signal real-time processing system based on multi-core distributed processing described in the present invention, wherein: the signal input unit is used to identify the encoding protocol of the input video digital signal, and receive the input video digital signal through the external interface of the preset multi-communication protocol, and send the input video digital signal to the signal buffer unit; The external interfaces include HDMI interface, SDI interface, PCIE interface and USB interface; The multiple communication protocols include H.264 / AVC protocol, H.265 / HEVC protocol, RTSP protocol, PCI Express protocol and Ethernet protocol.

[0007] As a preferred solution of the digital signal real-time processing system based on multi-core distributed processing described in the present invention, wherein: the signal buffer unit is used to process the input video digital signal through the first buffer block, the frame extraction intermediate block and the second buffer block to obtain the video frame sequence to be analyzed; The signal buffer unit includes a first buffer block, a frame extraction middle block and a second buffer block; The input video digital signal is directly written into the first buffer area, and the frame extraction intermediate block performs frame extraction operation according to the frame rate and timestamp of the input video digital signal in the first buffer area to obtain an initial video frame sequence. A write operation is performed on each frame data based on the time series order in the initial video frame sequence. When a conflict occurs in the write operation, the current video frame is discarded. When no conflict occurs in the write operation, the current video frame is written to the second buffer block. The second buffer block completes data reception and stores the data to obtain a video frame sequence to be analyzed.

[0008] As a preferred solution of the real-time digital signal processing system based on multi-core distributed processing described in the present invention, the color space conversion unit is used to perform color space conversion processing on the video frame sequence to be analyzed to obtain a grayscale image sequence, hue H component data and saturation S component data corresponding to the video frame sequence to be analyzed, and sequentially perform RGB color space to grayscale color space conversion processing on each video frame in the video frame sequence to be analyzed to obtain corresponding video frame grayscale image data; The RGB color space to HSV color space conversion process is performed on each video frame in the video frame sequence to be analyzed in turn to obtain the hue H component data and saturation S component data corresponding to each frame.

[0009] As a preferred solution of the digital signal real-time processing system based on multi-core distributed processing described in the present invention, wherein: the filtering enhancement unit is used to perform two-dimensional fast Fourier transform on the video frame grayscale image data to obtain an initial high-frequency component and an initial low-frequency component, perform double-layer filtering on the initial high-frequency component and the initial low-frequency component through a cascade filter group to obtain a first high-frequency sub-band, a second high-frequency sub-band and an output low-frequency sub-band, perform threshold screening and inverse transformation on the first high-frequency sub-band, the second high-frequency sub-band and the output low-frequency sub-band to obtain enhanced video frame grayscale image data, and perform pseudo-color mapping on the enhanced video frame grayscale image data to obtain an enhanced video frame image; The cascade filter group includes a first layer of low-pass filters, a first layer of high-pass filters, an eight-sector filter group, a second layer of low-pass filters, a second layer of high-pass filters and a four-sector filter group.

[0010] As a preferred solution of the digital signal real-time processing system based on multi-core distributed processing described in the present invention, the logic of performing double-layer filtering processing on the initial high-frequency component and the initial low-frequency component through the cascade filter group includes: The initial high-frequency component is processed by a first-layer high-pass filter and an eight-sector filter group to obtain a first high-frequency sub-band, the initial low-frequency component is decomposed by a first-layer low-pass filter to obtain a first high-frequency component and a first low-frequency component, the first high-frequency component is calculated by a second-layer high-pass filter and a four-sector filter group to obtain a second high-frequency sub-band, and the first low-frequency component is calculated by a second-layer low-pass filter to obtain an output low-frequency sub-band.

[0011] As a preferred solution of the digital signal real-time processing system based on multi-core distributed processing described in the present invention, the logic of performing threshold screening and inverse transformation processing on the first high-frequency sub-band, the second high-frequency sub-band and the output low-frequency sub-band to obtain enhanced video frame grayscale image data includes: The frequency band coefficients in the output low-frequency sub-band are multiplied by a preset first amplification weight coefficient to obtain an amplified output low-frequency sub-band, and the amplified output low-frequency sub-band and the initial low-frequency component are added to obtain a fused low-frequency component; All frequency band coefficients in the first high-frequency sub-band are screened by a preset first high-frequency threshold, coefficients greater than the first high-frequency threshold are assigned a value of 0 to obtain a first denoised high-frequency sub-band, and the frequency band coefficients in the first denoised high-frequency sub-band are multiplied by a preset second weight coefficient to obtain a first output high-frequency sub-band; All frequency band coefficients in the second high-frequency sub-band are screened by a preset second high-frequency threshold, coefficients greater than the second high-frequency threshold are assigned a value of 0 to obtain a second denoised high-frequency sub-band, and the frequency band coefficients in the second denoised high-frequency sub-band are multiplied by a preset third weight coefficient to obtain a second output high-frequency sub-band; Based on the fused low-frequency component, the first output high-frequency sub-band and the second output high-frequency sub-band, inverse Fourier transform processing is performed to obtain enhanced video frame grayscale image data.

[0012] As a preferred solution of the digital signal real-time processing system based on multi-core distributed processing described in the present invention, the target extraction unit is used to calculate the current frame motion feature map according to the current frame image and the previous frame image of the video frame sequence to be analyzed, and obtain the moving target positioning mask through binarization processing and morphological processing; Extract the current frame image and the previous frame image adjacent to the current frame image in the video frame sequence to be analyzed, perform inverse Fourier transform on the second high frequency subband corresponding to the current frame image to obtain the current frame detail image, perform inverse Fourier transform on the second high frequency subband corresponding to the previous frame image to obtain the previous frame detail image, perform difference calculation on the current frame detail image and the previous frame detail image to obtain the current frame difference detail image; The hue H component corresponding to the current frame image and the hue H component corresponding to the previous frame image are subtracted to obtain the current frame difference hue map; The saturation S component corresponding to the current frame image and the saturation S component corresponding to the previous frame image are subtracted to obtain the current frame difference saturation map; A current frame motion feature map is obtained by weighted addition of the preset current frame difference detail map, the current frame difference hue map and the current frame difference saturation map. The current frame motion feature map is binarized based on a preset motion feature threshold to obtain a motion feature region map. The motion feature region map is sequentially dilated and eroded to obtain a moving target positioning mask.

[0013] As a preferred solution of the real-time digital signal processing system based on multi-core distributed processing described in the present invention, the process scheduling module is used to monitor the resource utilization and fault conditions of the processors in the signal receiving module, the filtering enhancement unit and the motion target identification unit, and allocates and schedules the digital signal processing process based on the resource utilization and fault conditions of each processor according to the preset processing priority rules and the user instructions of the terminal interaction module.

[0014] As a preferred solution of the real-time digital signal processing system based on multi-core distributed processing described in the present invention, wherein: the terminal interaction module is used to receive the user's operation instruction data, set the system preset parameters based on the user's operation instruction data, connect and interact with the cloud server, and upload the enhanced video frame images and motion target positioning masks corresponding to each video frame in the video frame sequence to be analyzed to the cloud server.

[0015] The beneficial effects of the present invention are as follows: the video digital signal is efficiently processed in real time at the front end, and is suitable for processing video signals of a large range and for a long time, such as traffic condition monitoring and industrial monitoring. The input unit is configured with a variety of communication protocol interfaces, which is compatible with the access of a variety of devices, which is conducive to improving the versatility and scalability of the system and adapting to the needs of different scenarios. The signal cache unit avoids the cumulative error caused by the system continuously receiving a large number of signals for a long time through the partition block cache and write conflict handling mechanism, which is conducive to reducing the system failure caused by data confusion and ensuring the integrity and accuracy of the final video frame sequence to be analyzed. Through the two-dimensional fast Fourier transform combined with the double-layer filtering of the cascade filter group, the high-frequency and low-frequency information of the image can be effectively separated, the different frequency features can be accurately extracted, the image quality can be improved, the enhanced video frame grayscale image data can be made clearer and more accurate, and the enhanced video frame image with better visual effect can be obtained after pseudo-color mapping. Strengthening the overall detail sharpness and contrast of the image is conducive to making the image subject more prominent and the visual effect more vivid. The target extraction unit locates the moving target by the features of the detail map and the hue map saturation map of the adjacent frame image, which is conducive to further target tracking and behavior analysis processing at the back end. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A basic framework diagram of a digital signal real-time processing system based on multi-core distributed processing provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0018] Example 1, reference Figure 1 , is an embodiment of the present invention, and provides a digital signal real-time processing system based on multi-core distributed processing, including: a data receiving module, a parallel processing module, a process scheduling module and a terminal interaction module; The data receiving module includes a signal input unit and a signal buffer unit, which are used to identify the encoding protocol of the input video digital signal, receive the input video digital signal through an external interface of a preset multi-communication protocol, and process the input video digital signal through a first buffer block, a frame extraction intermediate block and a second buffer block to obtain a video frame sequence to be analyzed; The parallel processing module includes a color space conversion unit, a filtering enhancement unit and a target extraction unit, which are used to perform color space conversion processing on a video frame sequence to be analyzed to obtain a grayscale image sequence, hue H component data and saturation S component data corresponding to the video frame sequence to be analyzed, perform two-dimensional fast Fourier transform and double-layer filtering processing of a cascade filter group on the video frame grayscale image data to calculate enhanced video frame grayscale image data, perform pseudo-color mapping on the enhanced video frame grayscale image data to obtain an enhanced video frame image, calculate a current frame motion feature map according to a current frame image and a previous frame image of the video frame sequence to be analyzed, and obtain a moving target positioning mask through binarization processing and morphological processing; The process scheduling module is used to monitor the resource utilization and fault conditions of the processors in the signal receiving module, the filter enhancement unit and the moving target identification unit and to allocate and schedule the processing processes according to the preset processing priority rules and user instructions; The terminal interaction module is used to receive the user's operation instruction data and set the system preset parameters, and upload the enhanced video frame image and the moving target positioning mask corresponding to each video frame in the video frame sequence to be analyzed to the cloud server.

[0019] In this embodiment, the signal input unit is used to identify the encoding protocol of the input video digital signal, receive the input video digital signal through the external interface of the preset multi-communication protocol, and send the input video digital signal to the signal buffer unit; The external interfaces include HDMI interface, SDI interface, PCIE interface and USB interface; The multiple communication protocols include H.264 / AVC protocol, H.265 / HEVC protocol, RTSP protocol, PCI Express protocol and Ethernet protocol.

[0020] In this embodiment, the signal buffer unit is used to process the input video digital signal through the first buffer block, the frame extraction intermediate block and the second buffer block to obtain a video frame sequence to be analyzed; The signal buffer unit includes a first buffer block, a frame extraction middle block and a second buffer block; The input video digital signal is directly written into the first buffer area, and the frame extraction intermediate block performs frame extraction operation according to the frame rate and timestamp of the input video digital signal in the first buffer area to obtain an initial video frame sequence. A write operation is performed on each frame data based on the time series order in the initial video frame sequence. When a conflict occurs in the write operation, the current video frame is discarded. When no conflict occurs in the write operation, the current video frame is written to the second buffer block. The second buffer block completes data reception and stores the data to obtain a video frame sequence to be analyzed.

[0021] In this embodiment, the color space conversion unit is used to perform color space conversion processing on the video frame sequence to be analyzed to obtain a grayscale image sequence, hue H component data and saturation S component data corresponding to the video frame sequence to be analyzed, and sequentially perform RGB color space to grayscale color space conversion processing on each video frame in the video frame sequence to be analyzed to obtain corresponding video frame grayscale image data; The RGB color space to HSV color space conversion process is performed on each video frame in the video frame sequence to be analyzed in turn to obtain the hue H component data and saturation S component data corresponding to each frame.

[0022] In this embodiment, the filtering enhancement unit is used to perform a two-dimensional fast Fourier transform on the video frame grayscale image data to obtain an initial high-frequency component and an initial low-frequency component, perform a double-layer filtering process on the initial high-frequency component and the initial low-frequency component through a cascade filter group to obtain a first high-frequency sub-band, a second high-frequency sub-band and an output low-frequency sub-band, perform threshold screening and inverse transformation on the first high-frequency sub-band, the second high-frequency sub-band and the output low-frequency sub-band to obtain enhanced video frame grayscale image data, and perform pseudo-color mapping on the enhanced video frame grayscale image data to obtain an enhanced video frame image; The cascade filter group includes a first layer of low-pass filters, a first layer of high-pass filters, an eight-sector filter group, a second layer of low-pass filters, a second layer of high-pass filters and a four-sector filter group.

[0023] In this embodiment, the logic of performing double-layer filtering on the initial high-frequency component and the initial low-frequency component through the cascade filter bank includes: The initial high-frequency component is processed by a first-layer high-pass filter and an eight-sector filter group to obtain a first high-frequency sub-band, the initial low-frequency component is decomposed by a first-layer low-pass filter to obtain a first high-frequency component and a first low-frequency component, the first high-frequency component is calculated by a second-layer high-pass filter and a four-sector filter group to obtain a second high-frequency sub-band, and the first low-frequency component is calculated by a second-layer low-pass filter to obtain an output low-frequency sub-band.

[0024] In this embodiment, the logic of performing threshold screening and inverse transformation processing on the first high frequency sub-band, the second high frequency sub-band and the output low frequency sub-band to obtain enhanced video frame grayscale image data includes: The frequency band coefficients in the output low-frequency sub-band are multiplied by a preset first amplification weight coefficient to obtain an amplified output low-frequency sub-band, and the amplified output low-frequency sub-band and the initial low-frequency component are added to obtain a fused low-frequency component; All frequency band coefficients in the first high-frequency sub-band are screened by a preset first high-frequency threshold, coefficients greater than the first high-frequency threshold are assigned a value of 0 to obtain a first denoised high-frequency sub-band, and the frequency band coefficients in the first denoised high-frequency sub-band are multiplied by a preset second weight coefficient to obtain a first output high-frequency sub-band; All frequency band coefficients in the second high-frequency sub-band are screened by a preset second high-frequency threshold, coefficients greater than the second high-frequency threshold are assigned a value of 0 to obtain a second denoised high-frequency sub-band, and the frequency band coefficients in the second denoised high-frequency sub-band are multiplied by a preset third weight coefficient to obtain a second output high-frequency sub-band; Based on the fused low-frequency component, the first output high-frequency sub-band and the second output high-frequency sub-band, inverse Fourier transform processing is performed to obtain enhanced video frame grayscale image data.

[0025] In this embodiment, the target extraction unit is used to calculate the current frame motion feature map according to the current frame image and the previous frame image of the video frame sequence to be analyzed, and obtain the moving target positioning mask through binarization processing and morphological processing; Extract the current frame image and the previous frame image adjacent to the current frame image in the video frame sequence to be analyzed, perform inverse Fourier transform on the second high frequency subband corresponding to the current frame image to obtain the current frame detail image, perform inverse Fourier transform on the second high frequency subband corresponding to the previous frame image to obtain the previous frame detail image, perform difference calculation on the current frame detail image and the previous frame detail image to obtain the current frame difference detail image; The hue H component corresponding to the current frame image and the hue H component corresponding to the previous frame image are subtracted to obtain the current frame difference hue map; The saturation S component corresponding to the current frame image and the saturation S component corresponding to the previous frame image are subtracted to obtain the current frame difference saturation map; A current frame motion feature map is obtained by weighted addition of the preset current frame difference detail map, the current frame difference hue map and the current frame difference saturation map. The current frame motion feature map is binarized based on a preset motion feature threshold to obtain a motion feature region map. The motion feature region map is sequentially dilated and eroded to obtain a moving target positioning mask.

[0026] In this embodiment, the process scheduling module is used to monitor the resource utilization and fault conditions of the processors in the signal receiving module, the filtering enhancement unit and the motion target identification unit, and allocates and schedules the digital signal processing process based on the resource utilization and fault conditions of each processor according to the preset processing priority rules and the user instructions of the terminal interaction module.

[0027] In this embodiment, the terminal interaction module is used to receive the user's operation command data, set the system preset parameters based on the user's operation command data, connect and interact with the cloud server, and upload the enhanced video frame image and motion target positioning mask corresponding to each video frame in the video frame sequence to be analyzed to the cloud server.

[0028] Among them, the signal input unit is equipped with a variety of communication protocol interfaces, which is compatible with a variety of devices, which is conducive to improving the versatility and scalability of the system and adapting to the needs of different scenarios. The signal cache unit avoids the cumulative error caused by the system continuously receiving a large number of signals for a long time through the partition block cache and write conflict handling mechanism, which is conducive to reducing system failures caused by data confusion and ensuring the integrity and accuracy of the final video frame sequence to be analyzed. Through the two-dimensional fast Fourier transform combined with the double-layer filtering of the cascade filter group, it can effectively separate the high-frequency and low-frequency information of the image, accurately extract different frequency features, improve image quality, make the enhanced video frame grayscale data clearer and more accurate, and obtain enhanced video frame images with better visual effects after pseudo-color mapping. Strengthen the overall detail sharpness and contrast of the image, make the image subject more prominent, and make the visual effect more vivid. The target extraction unit locates the moving target by the features of the detail map, hue map and saturation map of adjacent frame images, which is conducive to target tracking and behavior analysis processing at the back end.

[0029] It should be understood by those skilled in the art that the embodiments of the present invention may provide methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program codes. Among them, the storage medium may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic memory, flash memory, magnetic disk or optical disk. These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0030] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A digital signal real-time processing system based on multi-core distributed processing, characterized in that: include: Data receiving module, parallel processing module, process scheduling module and terminal interaction module; The data receiving module includes a signal input unit and a signal buffer unit, which are used to identify the encoding protocol of the input video digital signal, receive the input video digital signal through an external interface of a preset multi-communication protocol, and process the input video digital signal through a first buffer block, a frame extraction intermediate block and a second buffer block to obtain a video frame sequence to be analyzed; The parallel processing module includes a color space conversion unit, a filtering enhancement unit and a target extraction unit, which are used to perform color space conversion processing on a video frame sequence to be analyzed to obtain a grayscale image sequence, hue H component data and saturation S component data corresponding to the video frame sequence to be analyzed, perform two-dimensional fast Fourier transform and double-layer filtering processing of a cascade filter group on the video frame grayscale image data to calculate enhanced video frame grayscale image data, perform pseudo-color mapping on the enhanced video frame grayscale image data to obtain an enhanced video frame image, calculate a current frame motion feature map according to a current frame image and a previous frame image of the video frame sequence to be analyzed, and obtain a moving target positioning mask through binarization processing and morphological processing; The process scheduling module is used to monitor the resource utilization and fault conditions of the processors in the signal receiving module, the filtering enhancement unit and the moving target identification unit and to allocate and schedule the processing processes according to the preset processing priority rules and user instructions; The terminal interaction module is used to receive the user's operation instruction data and set the system preset parameters, and upload the enhanced video frame image and the moving target positioning mask corresponding to each video frame in the video frame sequence to be analyzed to the cloud server.

2. A digital signal real-time processing system based on multi-core distributed processing as claimed in claim 1, characterized in that: The signal input unit is used to identify the encoding protocol of the input video digital signal, receive the input video digital signal through an external interface with a preset multi-communication protocol, and send the input video digital signal to the signal buffer unit; The external interfaces include HDMI interface, SDI interface, PCIE interface and USB interface; The multiple communication protocols include H.264 / AVC protocol, H.265 / HEVC protocol, RTSP protocol, PCI Express protocol and Ethernet protocol.

3. A digital signal real-time processing system based on multi-core distributed processing as claimed in claim 1, characterized in that: The signal buffer unit is used to process the input video digital signal through the first buffer block, the frame extraction intermediate block and the second buffer block to obtain a video frame sequence to be analyzed; The signal buffer unit includes a first buffer block, a frame extraction middle block and a second buffer block; The input video digital signal is directly written into the first buffer area, and the frame extraction intermediate block performs frame extraction operation according to the frame rate and timestamp of the input video digital signal in the first buffer area to obtain an initial video frame sequence. A write operation is performed on each frame data based on the time series order in the initial video frame sequence. When a conflict occurs in the write operation, the current video frame is discarded. When no conflict occurs in the write operation, the current video frame is written to the second buffer block. The second buffer block completes data reception and stores the data to obtain a video frame sequence to be analyzed.

4. A digital signal real-time processing system based on multi-core distributed processing as claimed in claim 1, characterized in that: The color space conversion unit is used to perform color space conversion processing on the video frame sequence to be analyzed to obtain a grayscale image sequence, hue H component data and saturation S component data corresponding to the video frame sequence to be analyzed, and sequentially perform RGB color space to grayscale color space conversion processing on each video frame in the video frame sequence to be analyzed to obtain corresponding video frame grayscale image data; The RGB color space to HSV color space conversion process is performed on each video frame in the video frame sequence to be analyzed in turn to obtain the hue H component data and saturation S component data corresponding to each frame.

5. The digital signal real-time processing system based on multi-core distributed processing according to claim 1, characterized in that: The filtering enhancement unit is used to perform two-dimensional fast Fourier transform on the video frame grayscale image data to obtain an initial high-frequency component and an initial low-frequency component, perform double-layer filtering on the initial high-frequency component and the initial low-frequency component through a cascade filter group to obtain a first high-frequency sub-band, a second high-frequency sub-band and an output low-frequency sub-band, perform threshold screening and inverse transformation on the first high-frequency sub-band, the second high-frequency sub-band and the output low-frequency sub-band to obtain enhanced video frame grayscale image data, and perform pseudo-color mapping on the enhanced video frame grayscale image data to obtain an enhanced video frame image; The cascade filter group includes a first layer of low-pass filters, a first layer of high-pass filters, an eight-sector filter group, a second layer of low-pass filters, a second layer of high-pass filters and a four-sector filter group.

6. A digital signal real-time processing system based on multi-core distributed processing as claimed in claim 5, characterized in that: The logic of double-layer filtering the initial high-frequency component and the initial low-frequency component by the cascaded filter bank includes: The initial high-frequency component is processed by a first-layer high-pass filter and an eight-sector filter group to obtain a first high-frequency sub-band, the initial low-frequency component is decomposed by a first-layer low-pass filter to obtain a first high-frequency component and a first low-frequency component, the first high-frequency component is calculated by a second-layer high-pass filter and a four-sector filter group to obtain a second high-frequency sub-band, and the first low-frequency component is calculated by a second-layer low-pass filter to obtain an output low-frequency sub-band.

7. A digital signal real-time processing system based on multi-core distributed processing as claimed in claim 5, characterized in that: The logic of performing threshold screening and inverse transformation processing on the first high frequency sub-band, the second high frequency sub-band and the output low frequency sub-band to obtain enhanced video frame grayscale image data includes: The frequency band coefficients in the output low-frequency sub-band are multiplied by a preset first amplification weight coefficient to obtain an amplified output low-frequency sub-band, and the amplified output low-frequency sub-band and the initial low-frequency component are added to obtain a fused low-frequency component; All frequency band coefficients in the first high-frequency sub-band are screened by a preset first high-frequency threshold, coefficients greater than the first high-frequency threshold are assigned a value of 0 to obtain a first denoised high-frequency sub-band, and the frequency band coefficients in the first denoised high-frequency sub-band are multiplied by a preset second weight coefficient to obtain a first output high-frequency sub-band; All frequency band coefficients in the second high-frequency sub-band are screened by a preset second high-frequency threshold, coefficients greater than the second high-frequency threshold are assigned a value of 0 to obtain a second denoised high-frequency sub-band, and the frequency band coefficients in the second denoised high-frequency sub-band are multiplied by a preset third weight coefficient to obtain a second output high-frequency sub-band; Based on the fused low-frequency component, the first output high-frequency sub-band and the second output high-frequency sub-band, inverse Fourier transform processing is performed to obtain enhanced video frame grayscale image data.

8. The digital signal real-time processing system based on multi-core distributed processing according to claim 1, characterized in that: The target extraction unit is used to calculate the current frame motion feature map according to the current frame image and the previous frame image of the video frame sequence to be analyzed, and obtain the moving target positioning mask through binarization processing and morphological processing; Extract the current frame image and the previous frame image adjacent to the current frame image in the video frame sequence to be analyzed, perform inverse Fourier transform on the second high frequency subband corresponding to the current frame image to obtain the current frame detail image, perform inverse Fourier transform on the second high frequency subband corresponding to the previous frame image to obtain the previous frame detail image, perform difference calculation on the current frame detail image and the previous frame detail image to obtain the current frame difference detail image; The hue H component corresponding to the current frame image and the hue H component corresponding to the previous frame image are subtracted to obtain the current frame difference hue map; The saturation S component corresponding to the current frame image and the saturation S component corresponding to the previous frame image are subtracted to obtain the current frame difference saturation map; A current frame motion feature map is obtained by weighted addition of the preset current frame difference detail map, the current frame difference hue map and the current frame difference saturation map. The current frame motion feature map is binarized based on a preset motion feature threshold to obtain a motion feature region map. The motion feature region map is sequentially dilated and eroded to obtain a moving target positioning mask.

9. The digital signal real-time processing system based on multi-core distributed processing according to claim 1, characterized in that: The process scheduling module is used to monitor the resource utilization and fault conditions of the processors in the signal receiving module, the filtering enhancement unit and the motion target identification unit, and allocates and schedules the digital signal processing process based on the resource utilization and fault conditions of each processor according to the preset processing priority rules and the user instructions of the terminal interaction module.

10. The digital signal real-time processing system based on multi-core distributed processing according to claim 1, characterized in that: The terminal interaction module is used to receive the user's operation instruction data, set the system preset parameters based on the user's operation instruction data, connect and interact with the cloud server, and upload the enhanced video frame image and motion target positioning mask corresponding to each video frame in the video frame sequence to be analyzed to the cloud server.

Images