Intelligent block image transmission system for picture processing

Abstract

The invention belongs to the technical field of image transmission and intelligent processing, and particularly relates to an intelligent block image transmission system for picture processing, which comprises the following steps: firstly, executing non-redundancy preprocessing on an input image to eliminate image interference and identify a key area; obtaining the texture complexity and criticality level of each region through sub-region division and feature calculation, so as to generate an image complexity-criticality map; taking the image complexity-criticality map as a basis, abandoning a fixed blocking mode, and making a differential blocking rule according to regional features; meanwhile, key degree, size and position labels are added to the blocks; based on this, the texture complexity and criticality of different regions of the image are extracted through image preprocessing, and then the differentiated blocking rule is formulated based on the features, so that the method has the advantages of realizing depth adaptation of the blocking strategy and the image content features, and avoiding detail loss of the key region and redundant data transmission of the non-key region caused by fixed blocking.

Description

Technical Field

[0001] This invention relates to the field of image transmission and intelligent processing technology, specifically to an intelligent segmented image transmission system for image processing. Background Technology

[0002] With the widespread adoption of high-definition image acquisition equipment (such as 4K industrial cameras, high-definition medical CT equipment, and drone aerial photography devices), the amount of image data is growing exponentially. The raw data size of a single 4K image can reach 10-20MB, and the bandwidth requirements for continuous transmission are extremely high. However, the transmission links in most practical application scenarios suffer from the problems of "limited bandwidth and poor stability": for example, IoT terminals in remote areas rely on narrowband communication (bandwidth is usually <1Mbps), and the wireless link during drone inspection is affected by airflow and electromagnetic interference, resulting in bandwidth fluctuations (2Mbps on sunny days, dropping to 500kbps on rainy days). Satellite communication links suffer from high latency and high packet loss rate (packet loss rate often reaches more than 15%).

[0003] To reduce transmission pressure, existing technologies generally adopt the "block transmission" approach, which involves dividing the image into several sub-blocks and compressing and transmitting them separately. However, the current solution has three major drawbacks, resulting in low transmission efficiency, poor image quality, and weak scene adaptability. Specific problems and examples are as follows:

[0004] 1. Fixed Block Segmentation Strategy, Unconsidered Image Content Features: Existing technologies often use preset fixed-size blocks (such as 64×64 or 128×128 pixel blocks), which cannot adapt to the differences in information density across different regions of the image. For example, in telemedicine scenarios, transmitting chest CT images reveals that lung lesions have high texture complexity (requiring high resolution to preserve details), while the chest wall is a solid-color area (high information redundancy). If a fixed 128×128 block is used, the lesion area will lose details due to the excessively large blocks (e.g., tiny calcifications cannot be identified), while the solid-color area will transmit a large amount of repetitive data due to the excessively small blocks (redundant data accounts for more than 35%), wasting bandwidth and affecting the accuracy of doctors' diagnoses.

[0005] 2. Disconnect between chunked transmission and link status, resulting in weak anti-interference capability: Existing systems only send chunked data according to preset rules, without real-time sensing of changes in link bandwidth and packet loss rate. This leads to severe packet loss when the link deteriorates and low bandwidth utilization when the link is idle. For example, in a drone-based farmland inspection scenario, the wireless link between the ground station and the drone is affected by cloud cover, causing the bandwidth to drop sharply from 1.5Mbps to 300kbps. If the existing system transmits according to the preset 256×256 chunks (8KB per chunk after compression) with the original 1.5Mbps bandwidth, the packet loss rate per frame reaches 40% when the link bandwidth is insufficient. The images received by the ground station show a large number of "missing chunks," making it impossible to fully identify areas of crop yellowing disease. Even after the link recovers, the system does not adjust the chunking parameters in time and continues to transmit at low resolution, missing the opportunity to improve image quality.

[0006] 3. Image preprocessing and chunked transmission are disconnected, lacking data support: Existing "image processing" (i.e., preprocessing without additional redundant data, such as noise reduction and normalization) is only used to optimize the image itself, without combining the processing results with the chunking strategy, leading to blind chunking decisions. For example, in security monitoring scenarios, when transmitting low-light images at night, the preprocessing stage has already identified "pedestrians" as critical areas and "walls" as non-critical areas through edge detection. However, the existing chunking system still uses a uniform 256×256 chunk, resulting in insufficient chunking of pedestrian areas (blurred edges, inability to identify clothing features) and redundant chunking of wall areas (increasing transmission time by 20%), failing to meet the monitoring requirements for clear tracking of key targets. Summary of the Invention

[0007] To address the technical problems mentioned above, such as a fixed segmentation strategy, lack of integration with image content features, disconnect between segmented transmission and link status, weak anti-interference capability, disconnect between image preprocessing and segmented transmission, and lack of data support, this invention provides the following technical solution:

[0008] An intelligent segmented image transmission system for image processing, comprising:

[0009] The image preprocessing and feature extraction module is used to perform non-redundant preprocessing on the input image to eliminate image interference and identify key regions; and then obtain the texture complexity and keyness level of each region through sub-region division and feature calculation to generate an image complexity-keyness map.

[0010] The content-aware dynamic segmentation module is used to abandon the fixed segmentation mode based on the image complexity-keyness map and formulate differentiated segmentation rules according to regional features; at the same time, keyness, size and location labels are added to the segments.

[0011] The real-time link status monitoring module is responsible for capturing the transmission link status in real time. First, it calculates key link parameters by periodically sending probe packets and combining the latency and packet loss data fed back by the receiving end. Then, it quantifies the link status into three levels: excellent, good, and poor.

[0012] The link-linked transmission control module is used to establish a dynamic mapping relationship based on both link status and block criticality, so as to adjust the compression rate of blocks with different criticalities, plan the sending priority, and update parameters in real time as the link status changes.

[0013] The block redundancy elimination and intelligent retransmission module is used to first calculate the similarity of adjacent low-criticality blocks, and only transmit the reference block and the difference data to eliminate redundancy; then, based on the feedback from the receiving end, it prioritizes the retransmission of critical blocks, and recovers non-critical lost blocks through interpolation.

[0014] The block quality verification and feedback module is used to first quantitatively evaluate the quality, use PSNR to determine compression distortion for key blocks, and use SSIM to detect integrity of the overall image after stitching, accurately locating substandard blocks and areas with stitching deviations; then, dynamic feedback optimization is performed, feeding back quality issues to the sending end in real time to trigger parameter adjustments.

[0015] As a preferred embodiment of the intelligent segmented image transmission system for image processing described in this invention, the image preprocessing and feature extraction module includes:

[0016] The image preprocessing unit is used to preprocess the input image, performing grayscale normalization, Gaussian noise reduction, and semantic segmentation in sequence.

[0017] The feature calculation unit is used to divide the preprocessed image into 16×16 pixel sub-regions and calculate the texture complexity and keyness level of each sub-region to generate an image complexity-keyness map.

[0018] As a preferred embodiment of the intelligent segmented image transmission system for image processing described in this invention, the content-aware dynamic segmentation module includes:

[0019] The chunking rule formulation unit is used to formulate chunking rules based on the image complexity-keyness map.

[0020] For highly complex and critical regions: the block size is set to 64×64 pixels to minimize the loss of detail;

[0021] Medium complexity and medium criticality region: block size set to 128×128 pixels;

[0022] Low complexity and low criticality areas: block size is set to 256×256-512×512 pixels, and adjacent areas of the same level are merged into a super-large block (such as 1024×1024) to reduce the number of blocks;

[0023] The segmentation marking unit is used to add key labels, size labels, and position coordinates to each segment to facilitate subsequent transmission control.

[0024] As a preferred embodiment of the intelligent segmented image transmission system for image processing described in this invention, the real-time link status monitoring module includes:

[0025] The probe packet sending unit is used to periodically send link probe packets and enable the receiving end to report the reception delay and packet loss of the probe packets.

[0026] The link parameter calculation unit is used to calculate real-time link parameters: bandwidth = (probe packet size × number of successfully received packets) / probe period, packet loss rate = (number of lost probe packets / total number of sent packets) × 100%;

[0027] The link classification unit is used to classify the link status into excellent, good, and poor levels.

[0028] As a preferred embodiment of the intelligent segmented image transmission system for image processing described in this invention, the link-linked transmission control module includes:

[0029] The link-block mapping unit is used to establish the link state-block transmission parameter mapping relationship:

[0030] Link optimization: Set the compression rate of high-criticality blocks to 50% and the compression rate of low-criticality blocks to 70%, and send them in descending order of block criticality;

[0031] Good link: Set the compression rate of high-criticality blocks to 60%, and the compression rate of low-criticality blocks to 80%. Prioritize sending blocks of criticality levels 1-3, and supplement with blocks of level 4-5 during idle periods.

[0032] Poor link performance: The compression rate of high-criticality blocks is set to 70%, and the compression rate of low-criticality blocks is set to 90%. Only blocks of criticality level 1-2 are sent. Non-critical blocks are temporarily stored locally and resent after the link is restored.

[0033] The dynamic adjustment unit is used to update the block compression rate and transmission priority in real time when the link status changes, so as to avoid packet loss and bandwidth waste.

[0034] As a preferred embodiment of the intelligent block image transmission system for image processing described in this invention, the block redundancy elimination and intelligent retransmission module includes:

[0035] The redundancy elimination unit is used to calculate pixel differences between adjacent low-key blocks. If the similarity is >90%, only the baseline block and the difference data are transmitted to reduce redundancy.

[0036] The intelligent retransmission unit prioritizes retransmitting criticality level 1-2 blocks when the receiver returns a list of lost blocks. For lost level 4-5 blocks, the receiver recovers them using a neighboring block interpolation algorithm without retransmission, thus reducing link occupancy.

[0037] As a preferred embodiment of the intelligent segmented image transmission system for image processing described in this invention, the segmented quality verification and feedback module includes:

[0038] The segment quality assessment unit is used to calculate quality indicators for each received segment:

[0039] Key segments: PSNR is used to evaluate compression distortion. If PSNR < scene preset threshold, the segment is marked as substandard.

[0040] Overall image after stitching: SSIM is used to evaluate the integrity of the stitched blocks. If SSIM < 0.85, the block area with stitching deviation is located.

[0041] The quality feedback unit is used to provide real-time feedback of the list of substandard blocks and splicing deviation areas to the link-linked transmission control module at the sending end, triggering secondary optimization.

[0042] Compared with existing technologies:

[0043] The "image complexity-criticality map" generated by the image preprocessing and feature extraction modules provides precise data support for the content-aware dynamic segmentation module, enabling it to break free from fixed segmentation patterns and formulate differentiated segmentation strategies based on image content features. Simultaneously, the link parameters captured by the real-time link status monitoring module guide the link-linked transmission control module to dynamically adjust the segmentation compression rate and transmission priority, achieving real-time adaptation between segmented transmission and link status. Through the collaborative efforts of these modules, the segmentation strategy deeply aligns with image content features, avoiding the loss of details in critical areas and redundant transmission in non-critical areas. Furthermore, the close linkage between segmented transmission and link status enhances anti-interference capabilities. Moreover, the deep integration of preprocessing and segmented transmission eliminates decision-making blindness, forming a closed-loop collaborative advantage from image feature extraction to segmented transmission control. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the overall framework of the present invention;

[0045] Figure 2 This is a schematic diagram of the image preprocessing and feature extraction module framework of the present invention;

[0046] Figure 3 This is a schematic diagram of the content-aware dynamic segmentation module framework of the present invention;

[0047] Figure 4 This is a schematic diagram of the link status real-time monitoring module framework of the present invention;

[0048] Figure 5 This is a schematic diagram of the link linkage transmission control module framework of the present invention;

[0049] Figure 6 This is a schematic diagram of the block redundancy elimination and intelligent retransmission module framework of the present invention;

[0050] Figure 7 This is a schematic diagram of the block quality verification and feedback module framework of the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0052] This invention provides an intelligent segmented image transmission system for image processing, such as... Figure 1 As shown, it includes:

[0053] The image preprocessing and feature extraction module is used to perform non-redundant preprocessing on the input image to eliminate image interference and identify key regions; and then obtain the texture complexity and keyness level of each region through sub-region division and feature calculation to generate an image complexity-keyness map.

[0054] The content-aware dynamic segmentation module is used to abandon the fixed segmentation mode based on the image complexity-keyness map and formulate differentiated segmentation rules according to regional features; at the same time, keyness, size and location labels are added to the segments.

[0055] The real-time link status monitoring module is responsible for capturing the transmission link status in real time. First, it calculates key link parameters by periodically sending probe packets and combining the latency and packet loss data fed back by the receiving end. Then, it quantifies the link status into three levels: excellent, good, and poor.

[0056] The link-linked transmission control module is used to establish a dynamic mapping relationship based on both link status and block criticality, so as to adjust the compression rate of blocks with different criticalities, plan the sending priority, and update parameters in real time as the link status changes.

[0057] The block redundancy elimination and intelligent retransmission module is used to first calculate the similarity of adjacent low-criticality blocks, and only transmit the reference block and the difference data to eliminate redundancy; then, based on the feedback from the receiving end, it prioritizes the retransmission of critical blocks, and recovers non-critical lost blocks through interpolation.

[0058] The block quality verification and feedback module is used to first quantitatively evaluate the quality, use PSNR to determine compression distortion for key blocks, and use SSIM to detect integrity of the overall image after stitching, accurately locating substandard blocks and areas with stitching deviations; then, dynamic feedback optimization is performed, feeding back quality issues to the sending end in real time to trigger parameter adjustments.

[0059] like Figure 2 As shown, the image preprocessing and feature extraction module includes:

[0060] The image preprocessing unit is used to preprocess the input image, performing grayscale normalization (eliminating illumination differences), Gaussian noise reduction (removing high-frequency noise), and semantic segmentation (based on a lightweight CNN model, such as MobileNet-UNet, to identify key regions, such as lesions in medical images and targets in surveillance images).

[0061] The feature calculation unit is used to divide the preprocessed image into 16×16 pixel sub-regions and calculate the texture complexity (contrast and entropy values ​​are calculated through the gray-level co-occurrence matrix) and keyness level (key regions are set to level 1-3, and non-key regions are set to level 4-5) of each sub-region to generate an image complexity-keyness map.

[0062] like Figure 3 As shown, the content-aware dynamic segmentation module includes:

[0063] The chunking rule formulation unit is used to formulate chunking rules based on the image complexity-keyness map.

[0064] For highly complex and critical regions (such as CT lesions with a complexity > 0.8 and a criticality level of 1): the block size is set to 64×64 pixels to minimize the loss of detail;

[0065] Medium complexity and medium criticality regions (such as non-lesion areas of the lungs on CT scans, with a complexity of 0.4-0.8 and a criticality level of 2): the block size is set to 128×128 pixels;

[0066] For low-complexity and low-criticality regions (such as the CT chest wall, with a complexity of <0.4 and a criticality of level 5): the block size is set to 256×256-512×512 pixels, and adjacent regions of the same level are merged into a super-large block (such as 1024×1024) to reduce the number of blocks.

[0067] The segmentation marking unit is used to add key labels, size labels, and position coordinates to each segment to facilitate subsequent transmission control.

[0068] like Figure 4 As shown, the real-time link status monitoring module includes:

[0069] The probe packet sending unit is used to periodically send link probe packets (once every 100ms) and enable the receiving end to report the reception delay and packet loss of the probe packets;

[0070] The link parameter calculation unit is used to calculate real-time link parameters: bandwidth = (probe packet size × number of successfully received packets) / probe period, packet loss rate = (number of lost probe packets / total number of sent packets) × 100%;

[0071] The link classification unit is used to classify the link status into excellent (bandwidth > 2Mbps, packet loss rate < 5%), good (bandwidth 1-2Mbps, packet loss rate 5%-15%), and poor (bandwidth < 1Mbps, packet loss rate > 15%) levels.

[0072] like Figure 5 As shown, the link-linked transmission control module includes:

[0073] The link-block mapping unit is used to establish the link state-block transmission parameter mapping relationship:

[0074] Link optimization: Set the compression rate of high-criticality blocks to 50% (low compression, quality assurance), and set the compression rate of low-criticality blocks to 70%, sending blocks from high to low criticality.

[0075] Good link: Set the compression rate of high-criticality blocks to 60%, and the compression rate of low-criticality blocks to 80%. Prioritize sending blocks of criticality levels 1-3, and supplement with blocks of level 4-5 during idle periods.

[0076] Poor link performance: Set the compression rate of high-criticality blocks to 70% (to improve compression rate and ensure transmission), and set the compression rate of low-criticality blocks to 90%. Only send blocks of criticality level 1-2, and temporarily store non-critical blocks locally until the link is restored.

[0077] The dynamic adjustment unit is used to update the block compression rate and transmission priority in real time when the link status changes, so as to avoid packet loss and bandwidth waste.

[0078] like Figure 6 As shown, the block redundancy elimination and intelligent retransmission module includes:

[0079] The redundancy elimination unit is used to calculate pixel differences between adjacent low-key blocks (key level 4-5). If the similarity is >90% (determined by the SSIM algorithm), only the reference block and the difference data (such as the difference matrix) are transmitted to reduce redundancy.

[0080] The intelligent retransmission unit prioritizes retransmitting blocks of criticality level 1-2 when the receiver returns a list of lost blocks. For lost blocks of level 4-5, the receiver recovers them using a neighboring block interpolation algorithm (such as bilinear interpolation) without retransmission, reducing link usage. Specifically, for lost blocks with low criticality (level 4-5) and low texture complexity (smooth regions), the neighboring block interpolation algorithm is used for recovery, reducing link usage. If a lost block has low criticality but high texture complexity (including edges and complex structures), it is still marked as a 'block that needs to be retransmitted' even if the criticality is low, and it is retransmitted first when the link is idle.

[0081] like Figure 7 As shown, the block quality verification and feedback module includes:

[0082] The segment quality assessment unit is used to calculate quality indicators for each received segment:

[0083] Key segments (levels 1-3): PSNR (Peak Signal-to-Noise Ratio) is used to evaluate compression distortion. If PSNR < scene preset threshold (e.g., 35dB in medical scenarios), it is marked as a segment with substandard quality.

[0084] Overall image after stitching: SSIM (structural similarity) is used to evaluate the integrity of the stitched blocks. If SSIM < 0.85 (default threshold, adjustable for different scenarios), the stitching deviation area is located.

[0085] The quality feedback unit is used to provide real-time feedback of the list of substandard blocks and splicing deviation areas to the link-linked transmission control module at the sending end, triggering secondary optimization.

[0086] If the PSNR of a critical segment is below standard: the sender reduces the compression rate of that segment (e.g., from 60% to 50%) and retransmits it;

[0087] If there is a splicing deviation: the sending end adjusts the precision of the position coordinate marking of the block in that area (e.g., refines the positioning from 16×16 pixels to 8×8 pixels) to avoid subsequent splicing deviations.

[0088] In practical use, the specific steps are as follows:

[0089] S1: Perform non-redundant preprocessing on the input image to eliminate image interference and identify key regions; then, through sub-region division and feature calculation, obtain the texture complexity and keyness level of each region to generate an image complexity-keyness map.

[0090] S2: Based on the image complexity-keyness map, the fixed block mode is abandoned, and differentiated block rules are formulated according to regional characteristics; at the same time, keyness, size and location labels are added to the blocks;

[0091] S3: Responsible for real-time capture of transmission link status; firstly, by periodically sending probe packets and combining the latency and packet loss data fed back by the receiver, it calculates key link parameters; and quantifies the link status into three levels: excellent, good, and poor.

[0092] S4: Based on both link status and block criticality, a dynamic mapping relationship is established to adjust the compression rate of blocks with different criticalities, plan the sending priority, and update parameters in real time as the link status changes.

[0093] S5: First, calculate the similarity of adjacent blocks with low criticality, and only transmit the reference block and the difference data to eliminate redundancy; then, based on the feedback from the receiver, prioritize the retransmission of critical blocks, and recover non-critical lost blocks through interpolation.

[0094] S6: First, quantitatively evaluate the quality, use PSNR to determine compression distortion for key blocks, and use SSIM to detect integrity of the overall image after stitching, accurately locating substandard blocks and stitching deviation areas; then, dynamically optimize by feeding back quality issues to the sending end in real time to trigger parameter adjustments.

[0095] Although the present invention has been described above with reference to embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of the invention. In particular, as long as there is no structural conflict, the features in the disclosed embodiments can be combined with each other in any manner. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, the present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. An intelligent segmented image transmission system for image processing, characterized in that, include: The image preprocessing and feature extraction module is used to perform non-redundant preprocessing on the input image to eliminate image interference and identify key regions; Then, through sub-region division and feature calculation, the texture complexity and criticality level of each region are obtained to generate an image complexity-criticality map; The content-aware dynamic segmentation module is used to abandon the fixed segmentation mode based on the image complexity-keyness map and formulate differentiated segmentation rules according to regional features; at the same time, keyness, size and location labels are added to the segments. The real-time link status monitoring module is responsible for capturing the transmission link status in real time. First, it calculates key link parameters by periodically sending probe packets and combining the latency and packet loss data fed back by the receiving end. Then, it quantifies the link status into three levels: excellent, good, and poor. The link-linked transmission control module is used to establish a dynamic mapping relationship based on both link status and block criticality, so as to adjust the compression rate of blocks with different criticalities, plan the sending priority, and update parameters in real time as the link status changes. The block redundancy elimination and intelligent retransmission module is used to first calculate the similarity of adjacent low-criticality blocks, and only transmit the reference block and the difference data to eliminate redundancy; then, based on the feedback from the receiving end, it prioritizes the retransmission of critical blocks, and recovers non-critical lost blocks through interpolation. The block quality verification and feedback module is used to first quantitatively evaluate the quality, use PSNR to determine compression distortion for key blocks, and use SSIM to detect integrity of the overall image after stitching, accurately locating substandard blocks and areas with stitching deviations; then, dynamic feedback optimization is performed, feeding back quality issues to the sending end in real time to trigger parameter adjustments.

2. The intelligent segmented image transmission system for image processing according to claim 1, characterized in that, The image preprocessing and feature extraction module includes: The image preprocessing unit is used to preprocess the input image, performing grayscale normalization, Gaussian noise reduction, and semantic segmentation in sequence. The feature calculation unit is used to divide the preprocessed image into 16×16 pixel sub-regions and calculate the texture complexity and keyness level of each sub-region to generate an image complexity-keyness map.

3. The intelligent segmented image transmission system for image processing according to claim 1, characterized in that, The content-aware dynamic segmentation module includes: The chunking rule formulation unit is used to formulate chunking rules based on the image complexity-keyness map. For highly complex and critical regions: the block size is set to 64×64 pixels to minimize the loss of detail; Medium complexity and medium criticality region: block size set to 128×128 pixels; Low complexity and low criticality areas: block size is set to 256×256-512×512 pixels, and adjacent areas of the same level are merged into a super-large block to reduce the number of blocks; The segmentation marking unit is used to add key labels, size labels, and position coordinates to each segment to facilitate subsequent transmission control.

4. The intelligent segmented image transmission system for image processing according to claim 1, characterized in that, The real-time link status monitoring module includes: The probe packet sending unit is used to periodically send link probe packets and enable the receiving end to report the reception delay and packet loss of the probe packets. The link parameter calculation unit is used to calculate real-time link parameters: bandwidth = (probe packet size × number of successfully received packets) / probe period, packet loss rate = (number of lost probe packets / total number of sent packets) × 100%; The link classification unit is used to classify the link status into excellent, good, and poor levels.

5. The intelligent segmented image transmission system for image processing according to claim 1, characterized in that, The link-linked transmission control module includes: The link-block mapping unit is used to establish the link state-block transmission parameter mapping relationship: Link optimization: Set the compression rate of high-criticality blocks to 50% and the compression rate of low-criticality blocks to 70%, and send them in descending order of block criticality; Good link: Set the compression rate of high-criticality blocks to 60%, and the compression rate of low-criticality blocks to 80%. Prioritize sending blocks of criticality levels 1-3, and supplement with blocks of level 4-5 during idle periods. Poor link performance: The compression rate of high-criticality blocks is set to 70%, and the compression rate of low-criticality blocks is set to 90%. Only blocks of criticality level 1-2 are sent. Non-critical blocks are temporarily stored locally and resent after the link is restored. The dynamic adjustment unit is used to update the block compression rate and transmission priority in real time when the link status changes, so as to avoid packet loss and bandwidth waste.

6. The intelligent segmented image transmission system for image processing according to claim 1, characterized in that, The block redundancy elimination and intelligent retransmission module includes: The redundancy elimination unit is used to calculate pixel differences between adjacent low-key blocks. If the similarity is >90%, only the baseline block and the difference data are transmitted to reduce redundancy. The intelligent retransmission unit prioritizes retransmitting criticality level 1-2 blocks when the receiver returns a list of lost blocks. For lost level 4-5 blocks, the receiver recovers them using a neighboring block interpolation algorithm without retransmission, thus reducing link occupancy.

7. The intelligent segmented image transmission system for image processing according to claim 1, characterized in that, The block quality verification and feedback module includes: The segment quality assessment unit is used to calculate quality indicators for each received segment: Key segments: PSNR is used to evaluate compression distortion. If PSNR < scene preset threshold, the segment is marked as substandard. Overall image after stitching: SSIM is used to evaluate the integrity of the stitched blocks. If SSIM < 0.85, the block area with stitching deviation is located. The quality feedback unit is used to provide real-time feedback of the list of substandard blocks and splicing deviation areas to the link-linked transmission control module at the sending end, triggering secondary optimization.

Images