A camera cloud platform storage system and method

Abstract

The application relates to the technical field of image data processing, in particular to a camera cloud platform storage system and method, the system comprising: a dynamic and static region division module which extracts the distribution region of motion pixels by comparing the change range of brightness and color based on continuous pictures collected by a camera, determines the range of the dynamic region for the distribution region of the motion pixels, simultaneously identifies and marks the remaining static region, and generates dynamic and static region division results. In the application, the dynamic and static characteristics of picture content are identified and distinguished, different processing modes are adopted for the static region and the dynamic region, frame similarity calculation is used in the static region, only key frames are reserved, repeated data storage is reduced, and storage resource consumption is reduced. The dynamic region is stored by extracting motion features and classification, the storage strategy is adjusted according to the motion frequency and change value, and the storage quality and resource occupation are effectively balanced.

Description

Technical Field

[0001] This invention relates to the field of image data processing technology, and in particular to a camera cloud platform storage system and method. Background Technology

[0002] Image data processing technology primarily revolves around the acquisition, transmission, storage, analysis, and processing of images. This field involves technologies such as computer vision, image compression, video encoding / decoding, and artificial intelligence algorithms. By processing image data, it helps achieve functions such as target detection, pattern recognition, feature extraction, data storage, and analysis. Image data processing has wide applications in various industries, including security monitoring, medical imaging, autonomous driving, smart cities, and intelligent manufacturing.

[0003] The camera cloud platform storage system can upload data collected by distributed camera devices to the cloud for centralized storage and management via the network. It supports real-time viewing, historical playback, intelligent analysis, and multi-terminal access. Its main applications include security monitoring, intelligent transportation, and industrial monitoring, providing users with convenient storage and efficient image data processing capabilities while reducing hardware costs and improving data security and availability.

[0004] Existing technologies in camera cloud platforms struggle to accurately identify the dynamic and static characteristics of video content, leading to the use of the same transmission and storage methods for both static and dynamic images, resulting in high data redundancy in static areas. For example, in monitoring scenarios with little change over extended periods, repeatedly storing complete frames consumes significant storage space, increasing system costs. For dynamic areas, existing technologies fail to categorize and process data based on motion frequency or range of change, causing confusion in the priority of transmission and storage for high-frequency dynamic images and inefficient resource allocation. In poor network conditions, the failure to adjust data transmission strategies in real time or record incomplete data transmission can lead to data loss or severe transmission delays in dynamic scenes, further reducing data integrity and user experience. For instance, in the face of sudden dynamic changes in monitoring footage, network congestion may prevent the timely transmission of key frames, impacting subsequent data analysis and judgment. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a camera cloud platform storage system and method.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a camera cloud platform storage system comprising:

[0007] The dynamic and static region segmentation module is based on continuous images captured by the camera. It compares the range of changes in brightness and color, extracts the distribution area of ​​moving pixels, determines the range of dynamic regions based on the distribution area of ​​moving pixels, and identifies and marks the remaining static regions to generate dynamic and static region segmentation results.

[0008] The image repetition assessment module extracts pixel values ​​from the static region based on the dynamic and static region division results, obtains histogram data of the static region, calculates the similarity between histogram data of the static region, marks the static region and repeated images according to the similarity, records the frequency and change value of the dynamic motion region, and generates image data of image repetition and dynamic features.

[0009] The partitioned storage optimization module classifies static areas according to inter-frame repetition and dynamic areas according to the frequency and change value of motion vectors based on the image data of the image repetition and dynamic features, and performs partitioned compression storage to obtain optimized regional storage records.

[0010] Based on the optimized regional storage records, the dynamic transmission and feedback module sorts the compressed data of dynamic and static regions according to the priority of the regional storage data, optimizes the transmission of image data, and generates the camera storage image transmission data recording results.

[0011] As a further aspect of the present invention, the step of obtaining the distribution area of ​​the moving pixels specifically includes:

[0012] Based on continuous image data captured by the camera, the brightness and color values ​​of the image are extracted according to each pixel of the image frame. The brightness and color changes of each pixel in the continuous frames are compared and analyzed to obtain the amplitude and direction information of the pixel changes.

[0013] Based on the change amplitude and direction information of the pixels, the set of pixels whose changes exceed the threshold is determined, the region where the pixels are located is extracted and marked, and the boundary of the dynamic region is divided to obtain the distribution area information of the moving pixels.

[0014] As a further aspect of the present invention, the steps for obtaining the dynamic and static region division results are specifically as follows:

[0015] Based on the distribution area information of the moving pixels, the densely distributed pixels are aggregated by analyzing the adjacency relationship of the pixels to form the dynamic region boundary, the sparsely distributed pixels are removed, and the static region is marked by combining the average brightness value and the inter-frame change frequency to generate preliminary division data of dynamic and static regions.

[0016] Based on the preliminary division data of the dynamic and static regions, the following formula is used:

[0017]

[0018] Calculate the difference P at the boundary between dynamic and static regions after smoothing, and process the boundary between dynamic and static regions according to the difference P to obtain the result of the division of dynamic and static regions.

[0019] Where n is the total number of pixels at the boundary between the dynamic and static regions, and B i S represents the grayscale value of the i-th dynamic boundary pixel. i This represents the grayscale value of the i-th static boundary pixel.

[0020] As a further aspect of the present invention, the step of obtaining the similarity between histograms specifically includes:

[0021] Based on the dynamic and static region division results, the brightness and color values ​​of all pixels are extracted from the static region. The pixels are grouped and counted according to brightness and color channels, and the number of pixels in each group is recorded to generate histogram data of the static region.

[0022] For the histogram data of the static region, the formula is used:

[0023]

[0024] Calculate the histogram similarity value G to obtain the similarity information between histograms;

[0025] Where j represents the j-th interval in the histogram, n′ is the total number of intervals in the histogram, and H Aj H is the height value of the histogram of frame A in the j-th interval. Bj It is the height value of the histogram of frame B in the j-th interval, min(H Aj H Bj ) is the minimum histogram height value of frame A and frame B in the j-th interval.

[0026] As a further aspect of the present invention, the step of obtaining image data on image repetition and dynamic features specifically includes:

[0027] Based on the similarity information between the histograms, the static regions are analyzed frame by frame to identify the similarity values. Static regions with similarity values ​​higher than the image segmentation threshold are identified as high similarity regions. The images to which the high similarity regions belong are identified as high repetition images, and the labeling information of high repetition images is generated.

[0028] Based on the motion vector data of moving pixels in the dynamic region, the frequency and changes of the moving region are statistically analyzed by considering the direction of motion, distance, and inter-frame brightness changes. The motion characteristics of each dynamic region are extracted, including the average frequency and range of change. Combined with the marking information of the high repetition image, image data of image repetition and dynamic characteristics are obtained.

[0029] As a further aspect of the present invention, the optimized regional storage record acquisition step specifically comprises:

[0030] Based on the image data with the aforementioned image repetition and dynamic features, static areas are classified according to the inter-frame repetition. High repetition areas and low repetition areas are determined according to the repetition. High repetition areas are extracted and their key frames are directly stored. At the same time, the boundary positions of the areas are recorded. For low repetition areas, pixel values ​​are extracted frame by frame and encoded and compressed according to the frame order to generate a partitioned storage result based on repetition.

[0031] Based on the motion vector data of dynamic regions, the regions are classified according to their motion frequency and change value. Regions with frequencies below a set frequency threshold are marked as low-change dynamic regions and their storage code rate is reduced. For dynamic regions with frequencies above the frequency threshold, their motion change features are extracted and combined with the location record. Combined with the partitioned storage results of the repetition, optimized region storage records are generated.

[0032] As a further aspect of the present invention, the step of obtaining the image transmission data recording result stored in the camera specifically comprises:

[0033] Based on the optimized region storage records, the formula is adopted:

[0034]

[0035] Calculate the dimensionless priority F representing the i′-th region. i′ This allows us to obtain regional priority information.

[0036] Among them, C i′ C is the data variation coefficient for the i′-th region. max R is the maximum value of the data variation coefficient across all regions. i′ R is the data update frequency of the i′-th region. max It is the maximum value of the data update frequency for all regions, L i′ L is the storage load factor of the i′-th region. max It is the maximum value of the storage load factor;

[0037] Based on the area priority information, the transmission capacity is allocated in real time according to the network bandwidth to obtain the camera's stored image transmission data recording results.

[0038] A camera cloud platform storage method, wherein the camera cloud platform storage method is executed based on the aforementioned camera cloud platform storage system, includes the following steps:

[0039] S1: Based on the image captured by the camera, extract the pixel brightness and color values ​​frame by frame, mark moving pixels by inter-frame comparison, aggregate the moving pixel region to define the boundary range, analyze the brightness change amplitude and frequency characteristics of non-moving regions to mark static regions, and generate dynamic and static region division results.

[0040] S2: Based on the dynamic and static region division results, extract the pixel brightness and color values ​​of the static region to generate histogram data, mark the high similarity region by comparing the histogram similarity values ​​frame by frame, extract the motion vector data from the dynamic region to count the frequency and change amplitude, and generate image repetition and dynamic feature image data.

[0041] S3: Based on the image repetition and dynamic feature image data, extract keyframes or compress and store them in static areas according to repetition, and identify and classify dynamic areas according to motion vector frequency and change amplitude to generate optimized area storage records.

[0042] S4: Based on the optimized regional storage record, dynamically adjust the regional priority sorting, transmit key data and feature points to high-priority regions, prioritize the transmission of dynamic region data, selectively transmit feature values ​​to low-priority static regions, and generate priority sorted image transmission data.

[0043] S5: Based on the image transmission data sorted by priority, and combined with the network bandwidth allocation transmission capacity, high-priority data is transmitted first, while low-priority areas are paused or the untransmitted data frame numbers and ranges are recorded, generating the camera stored image transmission data recording results.

[0044] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0045] In this invention, by identifying and distinguishing the dynamic and static characteristics of the image content, differentiated processing methods are applied to static and dynamic areas. In static areas, inter-frame similarity calculation is used to retain only keyframes, reducing redundant data storage and lowering storage resource consumption. For example, for scenes with no significant changes, redundant frames can be effectively eliminated, reducing storage pressure. Dynamic areas are processed by extracting motion features and classifying them for storage. The storage strategy is adjusted based on motion frequency and change values. High-frequency dynamic areas are stored in complete data form, while low-frequency dynamic areas are stored with reduced bitrate, effectively balancing storage quality and resource consumption. Furthermore, during transmission, data is prioritized according to the image content, with high-priority dynamic areas transmitted first to ensure real-time requirements, while only simplified feature values ​​are transmitted for low-priority static areas, reducing transmission traffic. In the event of network anomalies, the numbers and ranges of incompletely transmitted dynamic and static data are recorded to ensure the recoverability of subsequent transmissions and prevent data loss. Attached Figure Description

[0046] Figure 1 This is a system flowchart of the present invention;

[0047] Figure 2 This is a flowchart illustrating how the distribution area of ​​moving pixels is obtained according to the present invention.

[0048] Figure 3 This is a flowchart illustrating the process of obtaining dynamic and static region division results according to the present invention;

[0049] Figure 4 This is a flowchart illustrating the calculation of similarity between histograms in this invention;

[0050] Figure 5 This is a flowchart illustrating the process of acquiring image data with image repetition and dynamic features according to the present invention.

[0051] Figure 6 This is a flowchart illustrating how the optimized region storage record is obtained according to the present invention;

[0052] Figure 7 This is a flowchart illustrating the process of obtaining the recorded data of the image transmission stored by the camera according to the present invention. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0054] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0055] Please see Figure 1 The present invention provides a technical solution: a camera cloud platform storage system comprising.

[0056] The dynamic and static region segmentation module acquires the brightness and color values ​​of each pixel based on continuous images captured by the camera. It compares the range of brightness and color changes frame by frame, marks pixels whose changes exceed a preset image segmentation threshold, extracts the distribution area of ​​moving pixels, and determines the range of dynamic regions through boundary aggregation. For parts not marked as dynamic regions, it identifies and marks static regions by analyzing the average brightness within the frame and the frequency of changes between frames. It processes the boundary between dynamic and static regions through boundary smoothing calculation to generate dynamic and static region segmentation results.

[0057] The image repetition assessment module extracts pixel values ​​from static regions based on the dynamic and static region segmentation results. It constructs brightness and color histograms for each frame according to color channels, obtains histogram data of static regions, calculates the similarity between histogram data of static regions, and judges the inter-frame repetition by comparing the similarity value with the image segmentation threshold. According to the preset similarity reference value, static regions with a similarity higher than the preset similarity reference value are marked as static regions with high similarity and high repetition images. For dynamic regions, the frequency and change value of the dynamic region are recorded according to the motion vector and change range of moving pixels, generating image data of image repetition and dynamic features.

[0058] The partitioned storage optimization module classifies static regions based on frame repetition and dynamic features of image data, and classifies dynamic regions based on the frequency and change value of motion vectors. It implements partitioned compressed storage, including storing high-repetition frames directly as keyframes based on the frame repetition threshold of static regions and recording the position of high-repetition regions. For static regions in low-repetition images, it obtains the pixel values ​​of each frame and encodes and compresses them in frame order. For dynamic regions, it identifies dynamic regions with frequencies lower than a set frequency threshold based on the motion frequency and change value recorded by motion vectors, marks these regions as low-change dynamic regions, and reduces their storage bitrate. For dynamic regions with higher frequencies, it extracts the change value and range of motion vectors, combines them with the marked dynamic region positions, and fully records their motion features to obtain the optimized region storage record.

[0059] The dynamic transmission and feedback module, based on the optimized regional storage records, sorts the compressed data of dynamic and static regions according to the priority of the regional storage data. Combining the current network bandwidth of the camera cloud platform, it obtains the current transmission capacity in real time. Through priority scheduling, it selects the motion change data and key feature points of high-priority dynamic regions for priority transmission, reduces the transmission frequency of low-priority static region data, or only transmits histogram feature values. Under network abnormal conditions, it records the frame number and data range of the untransmitted dynamic or static regions, and generates the camera storage image transmission data recording results.

[0060] The results of dynamic and static region segmentation include the set of moving pixels, motion boundary range, and boundary smoothing data for dynamic regions; the average brightness distribution, inter-frame change frequency, and region labeling information for static regions; image data on image repetition and dynamic features including inter-frame repetition classification labels for static regions, similarity data and feature classification labels for high-repetition static regions; and motion frequency data, change value distribution, and frequency classification labels for dynamic regions. The optimized region storage records include keyframe indices and storage locations for high-repetition static regions, compressed storage data for low-repetition static regions, reduced bitrate storage records for low-frequency dynamic regions, and compressed motion change data and partitioning feature information for high-frequency dynamic regions. The camera image transmission data recording results include motion change transmission records and key feature point transmission data for high-priority dynamic regions, compressed data transmission records for low-priority static regions, and the unsent frame numbers, frame loss time ranges, and data tags for abnormal transmissions.

[0061] Please see Figure 2 The specific steps for obtaining the distribution area of ​​moving pixels are as follows:

[0062] Based on continuous image data captured by the camera, the brightness and color values ​​of the image are extracted according to each pixel of the image frame. The brightness and color changes of each pixel in the continuous frames are compared and analyzed to obtain the amplitude and direction information of the pixel changes.

[0063] Based on continuous images captured by a camera, the data is uploaded to a cloud platform for centralized storage via the network. OpenCV is used to process the image frames stored on the cloud platform, reading the brightness and color values ​​of each pixel from the stored image sequence. This is done by loading the image data and converting it into a pixel matrix. Brightness values ​​are extracted by converting the RGB image to grayscale, while color values ​​are directly obtained from the RGB channel matrix data. The cloud platform compares the brightness and color changes of each pixel in consecutive frames to obtain the magnitude and direction of these changes. For example, for each pixel in two consecutive frames, the absolute difference in its brightness value is calculated, and it is determined whether it exceeds a set threshold. For instance, for a pixel in frames A and B, if the brightness in frame A is 120 and the brightness in frame B is 110, then the magnitude of the change is |120-110| = 10. If this magnitude is greater than the threshold (e.g., 8), then the pixel is recorded as a point of change. By calculating the difference in pixel color values ​​between two consecutive frames, the direction is determined by the largest component of the difference in the changed RGB channels. For example, for the color values ​​(100, 120, 90) in frame A and (105, 125, 100) in frame B, the differences in the R, G, and B channels are calculated as |105-100|=5, |125-120|=5, and |100-90|=10, respectively. The direction of change is denoted as "increasing the B component". The magnitude and direction of the change are stored in a new matrix for analysis. Based on the results of each inter-frame comparison, the magnitude and direction results are uploaded to a cloud database for unified management. After processing, the system provides basic pixel data for subsequent marking of changed areas.

[0064] Based on the change amplitude and direction information of pixels, the set of pixels whose changes exceed the threshold is determined, the regions where these pixels are located are extracted and marked, and the boundaries of the dynamic regions are divided to obtain the distribution area information of moving pixels.

[0065] Pixels whose changes exceed a preset threshold are marked. The marked regions are extracted using the contour recognition method in OpenCV within a cloud platform processing environment. First, the set of pixels exceeding the threshold is determined by comparing matrix data. These sets form a binary marked image, which is uploaded to the cloud for further analysis of dynamic regions. The distribution range of pixels in the marked image is processed by extracting contours. The distribution of moving regions is determined by extracting boundaries, and K-means clustering is used to further classify these pixels. Different dynamic region boundaries are identified based on the distribution characteristics of the pixels. The generated boundary information is stored as region coordinate data. Simultaneously, the final distribution result of the dynamic regions is stored in a cloud file as an image, thus achieving accurate identification and segmentation of dynamic regions based on changing pixels.

[0066] Please see Figure 3The specific steps for obtaining the dynamic and static region division results are as follows:

[0067] Based on the distribution area information of moving pixels, the algorithm analyzes the adjacency relationship of pixels to aggregate densely distributed pixels to form dynamic region boundaries, removes sparsely distributed pixels, and combines the average brightness value and inter-frame change frequency to mark static regions, generating preliminary division data of dynamic and static regions.

[0068] First, spatial distribution analysis is performed on the extracted motion pixel data. Using a boundary aggregation method, sparsely distributed pixels are removed from concentrated regions. By detecting the brightness variation and adjacency relationships of pixels surrounding each pixel, pixels meeting the dense distribution criteria are integrated into complete dynamic regions, while the boundary positions are recorded. Pixels not classified as dynamic regions are processed independently. These pixels are selected based on intra-frame brightness mean analysis and continuous frame change frequency statistics, identifying pixels with stable brightness variations and frequencies below a specified threshold, and marked as static pixel regions. For example, a camera continuously stores image data at 25 frames per second on a cloud platform, and the distribution regions of motion pixels are extracted through data processing. Assume the motion pixel distribution data at a certain moment is stored as a two-dimensional array, where a value of "1" represents a motion pixel and a value of "0" represents a non-motion pixel. When performing boundary aggregation on these motion pixels, the number of "1"s around each motion pixel is first calculated based on adjacency relationships to determine if the pixel belongs to a dense distribution region. If more than 50% of the pixels in a motion pixel and its eight neighboring regions are motion pixels, it is classified into a dynamic region. Using aggregation algorithms (such as connectivity analysis), adjacent dynamic regions are merged to form complete region boundaries. Boundary data is further stored as polygon coordinates for subsequent image analysis and segmentation tasks. For parts not marked as dynamic regions (pixels with a value of "0"), the intra-frame mean brightness and inter-frame brightness change frequency are analyzed point-by-point. The specific operations are as follows: Mean brightness: Extract the grayscale value of each pixel from the original frame data, calculate the average brightness of its corresponding small region (e.g., a 3×3 window), and determine whether the pixel is within a stable brightness range (e.g., fluctuation less than ±5%). Inter-frame change frequency: By comparing the brightness value change records of the pixel in consecutive frames, count the number of brightness value changes within a specified time window (e.g., within 1 second). If the change frequency is lower than a threshold (e.g., 2 times / second), it is determined to be a static region. Pixels meeting the above conditions are marked as static regions, and the region segmentation result file stored in the cloud platform is updated.

[0069] Based on the preliminary division data of dynamic and static regions, the following formula is used:

[0070]

[0071] Calculate the difference P at the boundary between dynamic and static regions after smoothing, and process the boundary between dynamic and static regions according to the difference P to obtain the result of the division of dynamic and static regions.

[0072] Where P represents the brightness difference between dynamic and static pixels in the boundary region, n is the total number of pixels at the boundary between the dynamic and static regions, and B... i This represents the grayscale value of the i-th dynamic boundary pixel. The brightness value of the boundary pixel is extracted after image grayscale conversion. For example, brightness data can be directly obtained using the pixel intensity values ​​in the grayscale image. S i This represents the grayscale value of the i-th static boundary pixel. The brightness value of the corresponding boundary pixel is extracted from the image data after static region marking.

[0073] For example, if there are 4 pixels at the boundary between dynamic and static regions in an image frame, the set of pixel brightness values ​​for the dynamic region boundary is 120, 125, 118, 130, and the set of pixel brightness values ​​for the static region boundary is 122, 128, 117, 129. The calculation process is as follows:

[0074]

[0075]

[0076]

[0077] The results indicate that the brightness difference at the boundary between dynamic and static regions is 3.75 (relative brightness units). Comparing this to historical values ​​or a set baseline value (e.g., a baseline of 5), this low difference suggests a smoother brightness change at the boundary. Furthermore, boundary smoothing can optimize this result, ensuring a more continuous and natural division between dynamic and static regions, thereby improving the accuracy of region segmentation and visual appeal.

[0078] Please see Figure 4 The specific steps for calculating the similarity between histograms are as follows:

[0079] Based on the results of dynamic and static region segmentation, the brightness and color values ​​of all pixels are extracted from the static region. The pixels are grouped and counted according to brightness and color channels, and the number of pixels in each group is recorded to generate histogram data of the static region.

[0080] First, the stored frame image data is loaded from the cloud platform. Pixels marked as static areas are extracted, and the pixel values ​​of each channel within the static area are filtered according to the RGB color channels of the frame image. By statistically analyzing the brightness and color values ​​of each pixel, these values ​​are divided into different intensity intervals, and the number of pixels in each interval is recorded. The brightness histogram is generated by converting the grayscale values ​​of pixels within the static area from the original frame data. Each grayscale value is grouped and statistically analyzed to form a bar chart. The color histogram corresponds to each of the three RGB channels, and is generated by analyzing the intensity distribution of pixel values ​​within each channel individually. Finally, the brightness and color histograms of the static areas for each frame are stored in the cloud platform database as arrays, forming static area histogram data.

[0081] For histogram data of static regions, the formula is used:

[0082]

[0083] Calculate the histogram similarity value G to obtain the similarity information between histograms;

[0084] Where G represents the similarity between the brightness or a certain color channel histograms of two frames, with a value ranging from 0 to 1. The closer the value is to 1, the higher the similarity. j represents the j-th interval in the histogram, generated by grouping brightness or color values, such as dividing brightness values ​​into 0-15, 16-31, etc., or dividing the values ​​of each color channel into intervals to generate an index. n′ is the total number of intervals in the histogram, and H... Aj H is the height value of the histogram of frame A in the j-th interval. It is obtained by statistically analyzing the brightness or color channel values ​​of static regions in frame A. Bj It is the height value of the histogram of frame B in the j-th interval, obtained by statistically analyzing the brightness or color channel values ​​of static regions in frame B, min(H Aj H Bj ) is the minimum histogram height value of frame A and frame B in the j-th interval, representing the overlapping part of the pixel values ​​of the two frames in that interval.

[0085] For the luminance histogram: The static region luminance histogram data for frames A and B are divided into 4 intervals: 10, 20, 30, 40 and 15, 25, 20, 35.

[0086] Calculate the overlapping portion:

[0087] min(H A1 H B1 ) = min(10, 15) = 10

[0088] min(H A2 H B2 ) = min(20, 25) = 20

[0089] min(H A3 H B3 ) = min(30, 20) = 20

[0090] min(H A4 H B4 ) = min(40, 35) = 35

[0091] The similarity is calculated by summing the results:

[0092]

[0093] For use in color histograms (taking the R channel as an example):

[0094] The static region red channel histogram data of frames A and B are divided into 4 intervals: 8, 18, 28, 38 and 10, 20, 25, 30.

[0095] Calculate the overlapping portion:

[0096] min(H A1 H B1 ) = min(8, 10) = 8

[0097] min(H A2 H B2 ) = min(18, 20) = 18

[0098] min(H A3 H B3 ) = min(28, 25) = 25

[0099] min(H A4 H B4 ) = min(38, 30) = 30

[0100] The similarity is calculated by summing the results:

[0101]

[0102] The results showed that the similarity between the brightness histogram and the color histogram was 0.85 and 0.88, respectively.

[0103] Please see Figure 5 The specific steps for obtaining image data on image repetition and dynamic features are as follows:

[0104] Based on the similarity information between histograms, the similarity values ​​of static regions are analyzed frame by frame. Static regions with similarity values ​​higher than the image segmentation threshold are marked as high similarity regions, and the frames to which the high similarity regions belong are marked as high repetition frames, generating the marking information of high repetition frames.

[0105] The inter-frame repetition is determined by comparing similarity values ​​with an image segmentation threshold. The specific steps are as follows: First, the static region histogram data of the frames is loaded from the cloud platform. Similarity is calculated for both the luminance and color histograms, and the similarity value between each pair of frames is recorded. Next, the calculated similarity values ​​are compared with a preset image segmentation threshold (e.g., 0.9) to determine if the frames belong to a high-repetition category. For example, the similarity value of the luminance histograms of frames A and B is 0.85, which is lower than the image segmentation threshold of 0.9, therefore frame B is not marked as a high-repetition frame. Similarly, the similarity value of the red channel color histograms of frames C and D is 0.88, also lower than the threshold, so frame D is not marked as a high-repetition frame either. However, when the similarity of the blue channel color histograms of frames E and F is 0.92, which is higher than the threshold, frame F is marked as a high-repetition frame. All static region marking information with high similarity is integrated into a cloud-stored marking file as the basis for determining inter-frame repetition. Ultimately, the marked high-repetition images and highly similar static areas are used for further analysis and processing.

[0106] Based on the motion vector data of moving pixels in the dynamic region, the frequency and changes of the moving region are statistically analyzed by considering the direction of motion, distance, and inter-frame brightness changes. The motion characteristics of each dynamic region are extracted, including the average frequency and range of change. Combined with the marking information of high repetition images, image data of image repetition and dynamic characteristics are obtained.

[0107] For dynamic regions, the frequency and variation values ​​of the moving pixels are recorded by extracting their motion vectors and range of change. The specific steps are as follows: First, the position coordinates of the marked moving pixels in each frame's dynamic region are extracted, and the motion vectors, including direction and distance, are calculated by comparing consecutive frames. For example, if a pixel moves from position (10,15) to position (12,18) in a frame, its motion vector direction is 45 degrees, and the distance is 3 pixels. Then, the motion vectors of all pixels within the entire dynamic region are statistically analyzed to obtain the average motion frequency (e.g., 30 times per second) and motion variation value (e.g., brightness fluctuation range within ±20 gray levels). For example, in a set of dynamic regions, the average motion vector of pixels in a certain region is 5 pixels to the right per frame, with a motion frequency of 25 times per second; the brightness variation range is from 100 to 120 gray levels. After recording these statistical values, these data are integrated into an image data format representing image repetition and dynamic characteristics for subsequent storage and analysis. For example, regions with high frequency and large variation values ​​may represent fast-moving objects, while low-frequency, low-variation regions may represent slower background movement. Ultimately, the generated image data will be stored on a cloud platform, providing complete information on image repetition and dynamic characteristics.

[0108] Please see Figure 6The optimized steps for obtaining region storage records are as follows:

[0109] Based on image data with image repetition and dynamic features, static areas are classified according to inter-frame repetition. High-repetition areas and low-repetition areas are determined according to the repetition. High-repetition areas are extracted and their key frames are directly stored. At the same time, the boundary positions of the areas are recorded. For low-repetition areas, pixel values ​​are extracted frame by frame and encoded and compressed according to the frame order to generate a partitioned storage result based on repetition.

[0110] Based on image data stored on the camera cloud platform, which includes image repetition and dynamic features, static regions are classified according to inter-frame repetition. First, the static region histogram similarity and inter-frame repetition information for each frame are extracted. For example, in a set of consecutive frames, the static region similarity between frames A and B is 0.92, and the similarity between frames B and C is 0.89. Static regions with inter-frame similarity higher than a set repetition threshold (e.g., 0.9) are marked as high-repetition regions, and frame B is stored as a keyframe. The boundary positions and coordinates of the static regions are also recorded for subsequent processing. For frame C, since the similarity is lower than the repetition threshold, it is classified as a low-repetition image. Pixel values ​​for its static regions are extracted frame by frame, and the images are compressed and stored sequentially according to frame order using an encoding method (e.g., differential encoding). The encoding results are then stored in the cloud platform database. For example, assuming the static region repetition rates of frames 1 to 4 in 10 consecutive frames are 0.93, 0.91, 0.88, and 0.85 respectively, the static region of frames 1 to 2 is marked as high repetition, and frame 2 is stored as a keyframe. The coordinate range of the static region is recorded as (50, 50) - (200, 200). For the static region of frames 3 to 4, since its repetition rate is below the repetition threshold, pixel values ​​are extracted frame by frame, and the frame sequence data is encoded and stored using a compression algorithm. Finally, after classification, the storage information of the keyframes and compressed frames is synchronized to the cloud platform storage system, forming the optimized static region partitioning storage result.

[0111] Based on the motion vector data of dynamic regions, the regions are classified according to motion frequency and change value. Regions with frequencies below a set frequency threshold are marked as low-change dynamic regions and their storage code rate is reduced. For dynamic regions with frequencies above the frequency threshold, their motion change features are extracted and combined with position records. Combined with the partition storage results of repetition, optimized region storage records are generated.

[0112] Based on the dynamic feature data of the images stored on the camera cloud platform, dynamic regions are classified according to the motion frequency and change value recorded by motion vectors. First, the motion frequency of moving pixels in each dynamic region is extracted. For example, by comparing the motion vectors of pixels in consecutive frames, the motion frequency is calculated to be 20 times per second. At the same time, motion change values ​​are recorded, such as brightness changes within ±15 gray levels. Next, dynamic regions are classified according to a set frequency threshold (e.g., 15 times / second): if the motion frequency is lower than the frequency threshold, the region is marked as a low-change dynamic region, and the storage space requirement is reduced by adjusting the storage bitrate; if the motion frequency is higher than the frequency threshold, the change value and range of motion vectors are further extracted, and combined with the marked position of the dynamic region, the motion features are fully recorded. For example, suppose a frame's dynamic region is divided into two parts. Region A has a motion frequency of 10 times / second and a brightness variation range of ±10, which is lower than the set frequency threshold. Therefore, region A is marked as a low-change dynamic region, and the storage bitrate is adjusted to 70% of the standard value. Region B has a motion frequency of 25 times / second and a brightness variation range of ±25, which is higher than the frequency threshold. The direction and velocity change values ​​of its motion vector are extracted, and the marked position of region B within the frame is recorded as (30, 30)-(100, 100). Finally, the classified dynamic region storage records will be uploaded to the cloud platform database to form an optimized region storage result.

[0113] Please see Figure 7 The specific steps for obtaining the camera's stored image transmission data recording results are as follows:

[0114] Based on the optimized region storage records, the formula is used:

[0115]

[0116] Calculate the dimensionless priority F representing the i′-th region. i′ This allows us to obtain regional priority information.

[0117] Among them, F i′ Used for sorting dynamic and static regions, with higher priority regions being transmitted first. (C) i′ C is the data change coefficient for the i′-th region, used to measure the importance of the data. For dynamic regions, it is normalized based on the range of motion vector changes, for example, extracted and calculated through direction and velocity changes. For static regions, it is normalized based on the similarity difference of inter-frame histograms. max It is the maximum value of the data variation coefficient across all regions, used for dimensionless conversion, and is taken from the largest C value across all regions. i′ R i′R is the data update frequency of the i′-th region, used to represent the degree of dynamic change. For dynamic regions, it is obtained by statistically analyzing the update frequency of moving pixels in consecutive frames; for static regions, it is calculated by statistically analyzing the change frequency of the histogram between frames. max It is the maximum value of the data update frequency for all regions, representing the reference value for frequency normalization, L. i′ L is the storage load factor of the i′-th region, representing the data storage occupancy of the current region. It is derived by comprehensively considering the data volume, compression ratio, and current network bandwidth limitations. max It is the maximum value of the storage load factor, representing the normalized baseline value.

[0118] For example, if there are three regions (dynamic A, static B, and dynamic C), the parameters are as follows: data variation coefficient C. A =90, C B =50, C C =70, the maximum value of the data change coefficient is C max =90; Update frequency R A =20, R B =10, R C =15, maximum update frequency is R max =20; Storage load factor L A =1.5, L B =1.0, L C =1.2, maximum storage load factor is L max =1.5.

[0119] For dynamic region A:

[0120]

[0121] For static region B:

[0122]

[0123] For the dynamic region C:

[0124]

[0125] Based on the calculation results, the priority order is dynamic region A > dynamic region C > static region B. Dynamic region A has the highest priority because its data change coefficient and update frequency both reach their maximum values; dynamic region C has the second highest priority, and static region B has the lowest priority, making it suitable for low-frequency transmission or compression optimization. This result demonstrates that the dimensionless formula unifies different dimensional parameters, providing a unified standard for priority calculation between dynamic and static regions, thus optimizing the allocation of transmission resources.

[0126] Based on the regional priority information, the transmission capacity is allocated in real time according to the network bandwidth to obtain the recorded results of the camera's stored image transmission data.

[0127] Based on the current bandwidth of the camera cloud platform storage network, the current transmission capacity is obtained in real time. By reading the available capacity value of the current bandwidth, it is compared with the data volume requirements of dynamic area A, dynamic area C, and static area B. Priority is given to transmitting motion change data and key feature points of dynamic area A, allocating more bandwidth according to its priority value to ensure that the detailed information of dynamic changes is processed first. Next, the data content of dynamic area C is processed, and the transmission of secondary change features in this area is reduced according to the bandwidth allocation conditions. Finally, for the data of static area B, the transmission frequency is reduced, and only histogram feature values ​​are extracted for transmission. When the bandwidth capacity is insufficient, the transmission of static area B is suspended first, and the transmission frequency of dynamic area C is dynamically adjusted. The remaining bandwidth is concentrated on transmitting the key content of dynamic area A. In the event of bandwidth interruption or abnormality, the frame number and corresponding data range of dynamic or static areas that have not completed transmission are recorded. Finally, a camera storage image transmission data record result containing transmission records, incomplete areas, and untransmitted data prompts is generated.

[0128] A camera cloud platform storage method, which is executed based on the aforementioned camera cloud platform storage system, includes the following steps:

[0129] S1: Based on the image captured by the camera, extract the pixel brightness and color values ​​frame by frame, mark moving pixels by inter-frame comparison, aggregate the moving pixel region to define the boundary range, analyze the brightness change amplitude and frequency characteristics of non-moving regions to mark static regions, and generate dynamic and static region division results.

[0130] S2: Based on the results of dynamic and static region segmentation, extract the pixel brightness and color values ​​of static regions to generate histogram data. Mark high similarity regions by comparing histogram similarity values ​​frame by frame. Extract motion vector data from dynamic regions to count frequency and change amplitude, and generate image data of image repetition and dynamic features.

[0131] S3: Based on image data with repetition and dynamic features, keyframes are extracted or compressed and stored for static areas according to repetition, and dynamic areas are identified and classified according to motion vector frequency and change amplitude, generating optimized area storage records.

[0132] S4: Based on the optimized regional storage records, dynamically adjust the regional priority sorting, transmit key data and feature points to high-priority regions, prioritize the transmission of dynamic region data, selectively transmit feature values ​​to low-priority static regions, and generate priority sorted image transmission data.

[0133] S5: Based on priority sorting of image transmission data, combined with network bandwidth allocation of transmission capacity, high-priority data is transmitted first, and low-priority areas are paused or the untransmitted data frame number and range are recorded to generate the camera storage image transmission data recording results.

[0134] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A camera cloud platform storage system, characterized in that, The system includes: The dynamic and static region segmentation module is based on continuous images captured by the camera. It compares the range of changes in brightness and color, extracts the distribution area of ​​moving pixels, determines the range of dynamic regions based on the distribution area of ​​moving pixels, and identifies and marks the remaining static regions to generate dynamic and static region segmentation results. The image repetition assessment module extracts pixel values ​​from the static region based on the dynamic and static region division results, obtains histogram data of the static region, calculates the similarity between histogram data of the static region, marks the static region and repeated images according to the similarity, records the frequency and change value of the dynamic motion region, and generates image data of image repetition and dynamic features. The partitioned storage optimization module classifies static areas according to inter-frame repetition and dynamic areas according to the frequency and change value of motion vectors based on the image data of the image repetition and dynamic features, and performs partitioned compression storage to obtain optimized regional storage records. The specific steps for obtaining the optimized regional storage record are as follows: Based on the image data with the aforementioned image repetition and dynamic features, static areas are classified according to the inter-frame repetition. High repetition areas and low repetition areas are determined according to the repetition. High repetition areas are extracted and their key frames are directly stored. At the same time, the boundary positions of the areas are recorded. For low repetition areas, pixel values ​​are extracted frame by frame and encoded and compressed according to the frame order to generate a partitioned storage result based on repetition. Based on the motion vector data of dynamic regions, the regions are classified according to motion frequency and change value. Regions with frequencies below a set frequency threshold are marked as low-change dynamic regions and their storage code rate is reduced. For dynamic regions with frequencies above the frequency threshold, their motion change features are extracted and combined with position records. Combined with the partitioned storage results of the repetition, optimized region storage records are generated. Based on the optimized regional storage records, the dynamic transmission and feedback module sorts the compressed data of dynamic and static regions according to the priority of the regional storage data, optimizes the transmission of image data, and generates the camera storage image transmission data recording results.

2. The camera cloud platform storage system according to claim 1, characterized in that, The specific steps for obtaining the distribution area of ​​the moving pixels are as follows: Based on continuous image data captured by the camera, the brightness and color values ​​of the image are extracted according to each pixel of the image frame. The brightness and color changes of each pixel in the continuous frames are compared and analyzed to obtain the amplitude and direction information of the pixel changes. Based on the change amplitude and direction information of the pixels, the set of pixels whose changes exceed the threshold is determined, the region where the pixels are located is extracted and marked, and the boundary of the dynamic region is divided to obtain the distribution area information of the moving pixels.

3. The camera cloud platform storage system according to claim 2, characterized in that, The specific steps for obtaining the dynamic and static region division results are as follows: Based on the distribution area information of the moving pixels, the densely distributed pixels are aggregated by analyzing the adjacency relationship of the pixels to form the dynamic region boundary, the sparsely distributed pixels are removed, and the static region is marked by combining the average brightness value and the inter-frame change frequency to generate preliminary division data of dynamic and static regions. Based on the preliminary division data of the dynamic and static regions, the following formula is used: Calculate the difference at the boundary between dynamic and static regions after smoothing. According to the degree of difference The boundary between dynamic and static regions is processed to obtain the result of the dynamic and static region division. in, It is the total number of pixels at the boundary between dynamic and static regions. Indicates the first The grayscale value of a dynamic boundary pixel. Indicates the first The grayscale value of a static boundary pixel.

4. The camera cloud platform storage system according to claim 3, characterized in that, The specific steps for calculating the similarity between histograms are as follows: Based on the dynamic and static region division results, the brightness and color values ​​of all pixels are extracted from the static region. The pixels are grouped and counted according to brightness and color channels, and the number of pixels in each group is recorded to generate histogram data of the static region. For the histogram data of the static region, the formula is used: Calculate histogram similarity value This yields similarity information between histograms; in, Represents the first in the histogram Each interval It is the total number of intervals in the histogram. The histogram of frame A is in the th... The height value of each interval The histogram of frame B is in the th... The height value of each interval Are frames A and B in the 1st... The minimum value of the histogram height for each interval.

5. The camera cloud platform storage system according to claim 4, characterized in that, The specific steps for obtaining image data regarding image repetition and dynamic features are as follows: Based on the similarity information between the histograms, the static regions are analyzed frame by frame to identify the similarity values. Static regions with similarity values ​​higher than the image segmentation threshold are identified as high similarity regions. The images to which the high similarity regions belong are identified as high repetition images, and the labeling information of high repetition images is generated. Based on the motion vector data of moving pixels in the dynamic region, the frequency and changes of the moving region are statistically analyzed by considering the direction of motion, distance, and inter-frame brightness changes. The motion characteristics of each dynamic region are extracted, including the average frequency and range of change. Combined with the marking information of the high repetition image, image data of image repetition and dynamic characteristics are obtained.

6. The camera cloud platform storage system according to claim 1, characterized in that, The specific steps for obtaining the camera's stored image transmission data recording results are as follows: Based on the optimized region storage records, the formula is adopted: Calculate the representation of the first Dimensionless priority of each region This allows us to obtain regional priority information. in, It is the first The coefficient of change of data for each region It is the maximum value of the data change coefficient across all regions. It is the first Data update frequency for each region It is the maximum value of the data update frequency for all regions. It is the first Storage load factor of each region It is the maximum value of the storage load factor; Based on the area priority information, the transmission capacity is allocated in real time according to the network bandwidth to obtain the camera's stored image transmission data recording results.

7. A camera cloud platform storage method, characterized in that, The camera cloud platform storage system according to any one of claims 1-6 is executed, comprising the following steps: Based on the images captured by the camera, the pixel brightness and color values ​​are extracted frame by frame. Moving pixels are marked by inter-frame comparison, the moving pixel regions are aggregated to define the boundary range, and the brightness change amplitude and frequency characteristics of non-moving regions are analyzed to mark static regions, generating dynamic and static region division results. Based on the dynamic and static region division results, the pixel brightness and color values ​​of static regions are extracted to generate histogram data. High similarity regions are marked by comparing the histogram similarity values ​​frame by frame. Motion vector data is extracted from dynamic regions to statistically analyze frequency and change amplitude, thereby generating image repetition and dynamic feature image data. Based on the image repetition and dynamic feature data, key frames are extracted or compressed and stored for static areas according to repetition, and dynamic areas are identified and classified according to motion vector frequency and change amplitude, generating optimized area storage records. Based on the optimized regional storage records, the regional priority sorting is dynamically adjusted. Key data and feature points are transmitted to high-priority regions, dynamic regional data is transmitted first, and feature values ​​are selectively transmitted to low-priority static regions to generate priority sorted image transmission data. Based on the image transmission data sorted by priority, and combined with the network bandwidth allocation transmission capacity, high-priority data is transmitted first, while low-priority areas are paused or the untransmitted data frame numbers and ranges are recorded, generating the camera's stored image transmission data recording results.

Images