Image device data management method and system based on cloud-edge collaboration

Abstract

The present invention discloses an image device data management method and system based on cloud-edge collaboration, the method comprising: a cloud server configures task information and workflow information for an edge server and an edge device, and sends the task information and workflow information to the edge server; the edge server receives and parses the task information and workflow information, and sends a data acquisition instruction to the edge device; the edge device performs data acquisition according to the data acquisition instruction, obtains image data, and transmits the image data back to the edge server; the edge server processes the image data according to the workflow information, and uploads the processed image data to the cloud server; the cloud server receives and manages the image data according to a preset receiving mechanism, and constructs the received image data into an image device data packet. The present invention constructs an image device data packet based on cloud-edge collaboration, which can effectively improve the quality of image data.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an image device data management method and system based on cloud-edge collaboration. Background Art

[0002] At present, the rapid development of cloud computing, big data, and artificial intelligence technologies has provided strong support for enterprises to gradually shift from digitalization and automation to intelligence. In the process of transformation to intelligence, enterprises have continuously increased their investment in information construction and accumulated a large amount of data, such as structured data, audio and video data, and image data. As data assets are an important means of production, many companies are increasing their investment in data asset management, but at the same time, they are also facing many problems in data control during the management process.

[0003] Usually, the collection and management of image data generally includes obtaining a data stream, which may contain multiple frames of images. Therefore, single-frame image data can be obtained by extracting frames, and the obtained single-frame image is persistently stored. Although the above-mentioned data storage operation is relatively simple, there are also some problems. For example, in the frame extraction process, there are bad images due to problems with encoding and decoding. For example, the camera is not clear, the pixels are low, and other factors make the collected data unusable. In addition, since the scene of the picture captured by the fixed camera is single, there will be a large number of similar pictures, which makes the data source, scene, blood relationship, etc. unclear, and it is necessary to invest a lot of manpower in the subsequent process to manually screen and clean these data. Therefore, how to obtain high-quality image data is a problem that technicians in this field need to solve. Summary of the invention

[0004] The embodiments of the present invention provide an image device data management method and system based on cloud-edge collaboration, aiming to improve the quality of image data.

[0005] In a first aspect, an embodiment of the present invention provides an image device data management method based on cloud-edge collaboration, including:

[0006] In response to the user's selection instruction, the cloud server determines the corresponding edge server and edge device according to the selection instruction;

[0007] The cloud server configures task information and workflow information for the edge server and edge device, and sends the task information and workflow information to the edge server;

[0008] The edge server receives and parses the task information and workflow information, and then sends a data collection instruction to the edge device;

[0009] The edge device collects data according to the data collection instruction to obtain image data, and then transmits the image data back to the edge server;

[0010] The edge server processes the image data according to the workflow information, and uploads the processed image data to the cloud server;

[0011] The cloud server receives and manages the image data according to a preset receiving mechanism, and constructs the received image data into an image device data packet.

[0012] In a second aspect, an embodiment of the present invention provides an image device data management system based on cloud-edge collaboration, including a cloud server, an edge server, and an edge device, wherein:

[0013] The cloud server is used to respond to the user's selection instruction and determine the corresponding edge server and edge device according to the selection instruction; configure task information and workflow information for the edge server and edge device, and send the task information and workflow information to the edge server; receive and manage the image data according to a preset receiving mechanism, and construct the received image data into an image device data packet;

[0014] The edge server is used to receive and parse the task information and workflow information, and then send a data collection instruction to the edge device; process the image data according to the workflow information, and upload the processed image data to the cloud server;

[0015] The edge device is used to collect data according to the data collection instruction, obtain image data, and then transmit the image data back to the edge server.

[0016] The embodiment of the present invention provides an image device data management method and system based on cloud-edge collaboration. The embodiment of the present invention stores and manages the collected image data through cloud-edge collaboration, so that the data processing capability can be realized locally, avoiding delays in data transmission, thereby improving the business response speed, achieving the effect of quickly building image device data packets, and achieving more efficient management and application of image data. It can also reduce dependence on cloud servers, reduce network delays and security threats, and can also effectively reduce the bandwidth cost of data transmission on the network through local caching and other technologies. In addition, the embodiment of the present invention processes image data through edge servers to maximize the use of the computing power resources of edge servers and reduce the pressure on cloud servers to process data, so as to achieve the effect of building data assets with higher quality, easier management and traceability, thereby providing strong support for enterprises to build and manage data assets. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.

[0018] Figure 1 A flowchart of an image device data management method based on cloud-edge collaboration provided by an embodiment of the present invention;

[0019] Figure 2 A schematic diagram of the principle of an image device data management method based on cloud-edge collaboration provided by an embodiment of the present invention;

[0020] Figure 3 Another schematic diagram of the principle of an image device data management method based on cloud-edge collaboration provided by an embodiment of the present invention;

[0021] Figure 4 An exemplary schematic diagram of an image device data management method based on cloud-edge collaboration provided by an embodiment of the present invention;

[0022] Figure 5 Another exemplary schematic diagram of an image device data management method based on cloud-edge collaboration provided by an embodiment of the present invention;

[0023] Figure 6 A system architecture diagram of an image device data management system based on cloud-edge collaboration provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0026] It should also be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0027] It should be further understood that the term "and / or" used in the present description and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0028] See below Figure 1 , an embodiment of the present invention provides an image device data management method based on cloud-edge collaboration, specifically including: steps S101~S106.

[0029] Step S101: In response to a selection instruction from a user, the cloud server determines a corresponding edge server and edge device according to the selection instruction;

[0030] Step S102: The cloud server configures task information and workflow information for the edge server and edge device, and sends the task information and workflow information to the edge server;

[0031] Step S103: the edge server receives and parses the task information and workflow information, and then sends a data collection instruction to the edge device;

[0032] Step S104: the edge device collects data according to the data collection instruction to obtain image data, and then transmits the image data back to the edge server;

[0033] Step S105: the edge server processes the image data according to the workflow information, and uploads the processed image data to the cloud server;

[0034] Step S106: The cloud server receives and manages the image data according to a preset receiving mechanism, and constructs the received image data into an image device data packet.

[0035] This embodiment uses cloud (cloud server) and edge (edge ​​server, edge device, such as camera and other image acquisition devices) collaboration to build image device data packets to achieve storage and management of data assets. Figure 2 First, the cloud server selects the corresponding edge server and edge device (such as Figure 2The selected edge server and edge device are configured with corresponding task information and workflow information, and sent to the edge server. The edge server then sends data collection instructions to the edge device based on the task information and workflow information generated by the cloud server, such as Figure 3 As shown, the edge device collects image data and transmits it back to the edge server. The edge server processes the image data returned by the edge device based on the workflow information, and uploads it to the cloud server after processing. The cloud server receives the image data according to the set receiving mechanism, and thereby constructs the image data into an image device data packet.

[0036] This embodiment stores and manages the collected image data through cloud-edge collaboration, so that data processing capabilities can be realized locally, avoiding delays in data transmission, thereby improving business response speed, achieving the effect of quickly building image device data packets, and achieving more efficient management and application of image data. It can also reduce dependence on cloud servers, reduce network delays and security threats, and can also effectively reduce the bandwidth cost of data transmission on the network through local caching and other technologies. In addition, this embodiment processes image data through edge servers to maximize the use of edge server computing resources and reduce the pressure on cloud servers to process data, thereby achieving the effect of building data assets with higher quality, easier management and traceability, thereby providing strong support for enterprises to build and manage data assets.

[0037] In actual application scenarios, users can select a single edge server or multiple edge servers in the cloud server. Similarly, each edge server can correspond to a single edge device or multiple edge devices. The specific selection method can be set according to actual needs.

[0038] In one embodiment, the step S102 includes:

[0039] The cloud server sets task information according to the selection instruction, selects data processing components according to the selection instruction, and generates workflow information based on the data processing components; wherein the task information includes task execution frequency and task effective time; the workflow information includes data processing flow and data storage path;

[0040] Accordingly, the step S103 includes:

[0041] The edge server parses the task information to obtain the task execution frequency and the task effective time, and sends a data collection instruction to the edge device according to the task execution frequency and the task effective time;

[0042] The edge server parses the workflow information to obtain the data processing flow and data storage path, and performs data management according to the data processing flow and data storage path.

[0043] In this embodiment, after the user selects the edge server and the corresponding edge device in the cloud server, the cloud server will set the strategy for constructing the image device data packet accordingly, which may include quantity strategy, growth threshold strategy, and cycle strategy, and then select the workflow that the edge server needs to execute, and select data processing components, including data migration, similarity detection, quality detection, and data upload components. At the same time, the corresponding task information is set, such as setting the task execution frequency to single or timed multiple times, the task effective time, etc., and then the set task information and workflow information are sent to the specified edge server for execution.

[0044] After receiving the configuration tasks and workflow information sent by the cloud server, the edge server parses them, and then sends instructions to the corresponding edge devices according to the set frequency and rules, so that the edge devices start data collection, and store the collected data persistently in the specified path and rules configured in the workflow, so as to be persistently stored in the edge server. At the same time, the edge server starts to process the received workflow tasks. Specifically, first verify the timing information of the workflow, and then start to verify each data processing or data reporting component in the workflow, such as the quality detection component of the configuration data, the similarity deduplication component, and the cloud data upload component, so as to complete the execution of the workflow regularly according to the configuration.

[0045] In one embodiment, the data processing component includes a data quality detection component;

[0046] The step S105 comprises:

[0047] According to the following formula, the image data is subjected to mosaic detection using the data quality detection component:

[0048] ;

[0049] Wherein, r represents the result of mosaic detection, g represents the number of pixels of different gray levels in the image data, and s g represents the number of pixels of the largest first n gray levels, h represents the height of the image data, and w represents the width of the image data;

[0050] Comparing the mosaic detection result with a preset first threshold requirement;

[0051] If the result of the mosaic detection meets the preset first threshold requirement, retaining the corresponding image data;

[0052] If the result of the mosaic detection does not meet the preset first threshold requirement, the corresponding image data is deleted.

[0053] Furthermore, the step S105 further includes:

[0054] Performing Gaussian blur processing on the image data, and respectively calculating the gradient value of the image data before and after the Gaussian blur processing;

[0055] According to the following formula, the clarity of the image data is detected by using the data quality detection component:

[0056] ;

[0057] Where score represents the clarity score, N represents the total number of pixels in the image data, x represents the pixel index in the horizontal direction, y represents the pixel index in the vertical direction, and M represents the pixel index in the horizontal direction. O (x, y) represents the gradient value of the image data before Gaussian blur processing, M b (x, y) represents the gradient value of the image data after Gaussian blur processing;

[0058] comparing the clarity score with a preset second requirement threshold;

[0059] If the clarity score meets the preset second threshold requirement, retaining the corresponding image data;

[0060] If the clarity score does not meet the preset second threshold requirement, the corresponding image data is deleted.

[0061] In this embodiment, the data processing component may include a quality detection component, thereby realizing quality detection of the image data, thereby improving the quality of the image device data packet finally constructed by the cloud server. In this embodiment, when performing quality detection, two aspects may be included, one is mosaic detection, and the other is clarity detection.

[0062] When detecting whether the image data is a bad image (mosaic), first obtain the image data according to the image input path, and obtain the width w and height h of the image data. If the height or width of the image data exceeds the specified value, the image size is scaled proportionally. The image is then converted into a pixel matrix p[x, y] = (R, G, B), and a nested loop is used to traverse each pixel to check whether there is a missing area (that is, the values ​​of the three channels R, G, and B are equal). Next, the number of pixels at different gray levels g is calculated, and the proportion s of the number of gray pixels in the top n (for example, the top ten, because among the levels 0-255, the first ten levels have the highest proportion of pixels) is calculated to the total number of pixels. gAnd return the result. Then, the image data can be judged as a bad image according to the mosaic detection result, that is, when the mosaic detection result reaches the set threshold, it is judged as a bad image, otherwise, when the mosaic detection result does not reach the set threshold, it is judged as a normal image.

[0063] When performing clarity detection, the clarity of the upper and lower parts of the image data can be detected separately. Specifically, the image data can be read first, and the width w and height h of the image data can be obtained. If the height or width of the image data exceeds the specified value, the image size is scaled proportionally, and then the image data is divided into the upper half h1 and the lower half h2, and then the clarity scores s1 and s2 of the upper and lower halves are calculated respectively. That is, the image data is converted into a grayscale image, and the horizontal and vertical gradients s of the original image are calculated respectively. x and y , and use M O (x, y) records. Then apply Gaussian blur to the input grayscale image to reduce noise and details and smooth the image. Then calculate the horizontal and vertical gradients s of the blurred image bx and by , and use M b (x, y) records are recorded for comparison with the original image. The gradient difference between the original image and the blurred image is then calculated, and the gradient difference is used as the image clarity score indicator. And because the gradient of the clear image is large, and the gradient of the blurred image is small, it can be used to determine whether the clarity of the image data meets the standard. In practical applications, the clarity scores of the upper and lower parts can be calculated separately, and a clarity score list is returned. Then, according to the data in the clarity score list, it is determined whether the corresponding image data passes the clarity test.

[0064] In some optional embodiments, the mosaic detection result and the clarity scores of the upper and lower parts of the image are combined to evaluate the quality of the image data. If any one of the mosaic detection result and the clarity scores of the upper and lower parts of the image does not meet the corresponding threshold requirements, it is determined that the corresponding image data quality detection has failed, and the image data is deleted; if the mosaic detection result and the clarity scores of the upper and lower parts of the image both meet the corresponding threshold requirements, it is determined that the corresponding image data quality detection has passed, and the image data is retained.

[0065] In another embodiment, the data processing component further includes a similarity detection component;

[0066] The step S105 further includes:

[0067] The edge server uses a hash algorithm to calculate a hash mean for the image data;

[0068] Grouping and aggregating the image data according to the edge device ID to obtain multiple image data groups;

[0069] Based on the similarity detection component, similarity calculation is performed on the image data in each image data group, and deduplication processing is performed on the image data according to the result of the similarity calculation.

[0070] In this embodiment, the data processing component may also include a similarity detection component, thereby realizing similarity detection of image data, so that the image data can be deduplicated according to the similarity detection result, thereby not only improving the quality of the image data uploaded by the edge server, but also avoiding waste of server resources. Specifically, first, the image data whose hash mean has not been calculated in the database can be queried in a cyclic paging manner to obtain the storage address of the image data, and then the image data can be obtained according to the storage address, and the mean hash value of the image data can be calculated by the Hash algorithm. The function converts the image data into a grayscale image, and then divides the grayscale image into blocks, and then calculates the mean of each block, and generates a binary hash value based on the mean. The final hash value is a 64-bit (16×16) binary string, which is used to represent the content characteristics of the image, and then returns the calculated mean hash value, and updates the obtained hash mean into the database. Thereafter, all image data in the database (valid and not deduplicated) are queried, and grouped and aggregated according to the edge device ID, and then data deduplication is performed on each device. Specifically, the Hamming distance in the hash algorithm can be used. Calculate the similarity, where p represents the Hamming distance and h o and h i Represent two image data respectively. The smaller the Hamming distance, the higher the similarity. The images whose similarity reaches the set similarity threshold are deduplicated. In addition, for the hash mean of the image data under the same edge device, if the hash mean is the same, the previous image is retained, the database image status is updated, and the status of the deduplicated image is set to 1 to avoid subsequent comparison with other image data. The reason why the image data was deleted can also be recorded.

[0071] In a specific embodiment, after passing the above-mentioned quality detection and similarity detection, the edge server queries the image data whose status is processed, and uploads the image data to the cloud server for storage and management according to the upload path set in the workflow. At the same time, the edge server records the uploaded image information in the local database, and uploads the recorded information to the cloud server at the same time for subsequent data processing and better traceability.

[0072] In one embodiment, the step S106 includes:

[0073] The cloud server receives the image data uploaded by the edge server, and determines whether a preset receiving mechanism is triggered; wherein the preset receiving mechanism includes any one or more of a data growth rate threshold mechanism, a data cycle mechanism, and a data quantity mechanism;

[0074] If it is determined that the preset receiving mechanism is not triggered, continue to receive the image data;

[0075] If it is determined that the preset receiving mechanism is triggered, the reception of the image data is stopped, and a stop uploading instruction is sent to the edge server, so that the edge server stops uploading the image data.

[0076] Since there is an upper limit to the richness of samples collected by a single camera device, the cloud server in this embodiment receives and manages image data according to the set receiving mechanism to avoid insufficient data to meet the demand and to avoid waste of resources due to excessive data received. When the receiving mechanism is triggered, it stops receiving image data and sends a corresponding instruction to the edge server to stop uploading image data. When the receiving mechanism is not triggered, it can continue to receive image data until the receiving mechanism is triggered. Here, the receiving mechanism described in this embodiment can be any one or more of a data growth rate threshold mechanism, a data cycle mechanism and a data quantity mechanism. In other words, no matter which of the above mechanisms is triggered, it will be deemed that the receiving mechanism is triggered, thereby stopping receiving image data. Among them, the data growth rate threshold mechanism refers to using the data of the previous day as the benchmark data, calculating the daily growth rate, and stopping when the growth rate reaches the set threshold; the data cycle mechanism refers to receiving image data within a limited time period set to take effect, and stopping receiving if the time is reached; the data quantity mechanism refers to stopping receiving when the quantity reaches a certain value. Figure 4 As shown, this embodiment takes October 19, 2023 as the starting date, adopts the data growth rate threshold mechanism as the receiving mechanism, takes the data of the previous day as the benchmark data, calculates the daily growth rate, such as the new increase rate on October 20, 2023 is 60%, such as the new increase rate on October 21, 2023 is 25%, such as the new increase rate on October 22, 2023 is 10%, until the new increase rate of 202×××× is 1%, and stops receiving.

[0077] In actual applications, the cloud server can insert the received image data into the image data record table of the cloud server, and then the cloud server determines whether the receiving mechanism is triggered according to the edge device ID, thereby determining whether to stop receiving the image data. Preferably, when the image device data packet to be constructed is dynamically updated, the set receiving mechanism can be dynamically adjusted according to actual needs.

[0078] In one embodiment, after step S106, the following steps are included:

[0079] The cloud server uses the large model to perform regional label calibration on the image data in the image device data packet;

[0080] Use image data with region labels to train the specified model and / or algorithm.

[0081] Combination Figure 5 , after the image device data packet is constructed, the present embodiment divides the regional labels through the large model, so that it can be applied in the training of different model algorithms. Specifically, after the image device data packet is constructed, some pictures in the image device data packet are randomly extracted, and the regional label type is formulated, and then the extracted pictures are transmitted to the cloud large model, and the large model marks the regional label on the image device data packet; when a certain algorithm needs data for training, the required regional scenes can be screened out to obtain a set of regional data packets, and then the training data packet (including regional scenes and training requirements) can be constructed by adding the requirements description. After obtaining the training data packet, the algorithm can further pre-process the data (such as cleaning, pre-labeling, etc.), and then use it for algorithm training. At the same time, the training data packet can also be set to support version management, that is, it can be inherited and merged from any version, so that the training sample data can be more flexibly enriched, the personalized needs of the algorithm can be met, and the content-rich training data packet can be provided more quickly.

[0082] In addition, further, at the operational level, when a model is trained, it can be deployed to the production environment. The production environment will continue to receive camera capture data, and the deployed model will give annotations and alarm information. The image data that generates alarms will flow to the operation platform, and then the algorithm or operation personnel will select the alarm information data that the model judges as errors to build an operation data package. Combined with the training data package, secondary training and algorithm tuning are carried out to continuously optimize the algorithm and model, achieving the effect of building a "data-algorithm" closed loop.

[0083] like Figure 6 As shown, an embodiment of the present invention further provides an image device data management system based on cloud-edge collaboration, the system comprising a cloud server 100, an edge server 200 and an edge device 300, wherein:

[0084] The cloud server 100 is used to respond to the user's selection instruction and determine the corresponding edge server 200 and edge device 300 according to the selection instruction; configure task information and workflow information for the edge server 200 and edge device 300, and send the task information and workflow information to the edge server 200; receive and manage the image data according to a preset receiving mechanism, and construct the received image data into an image device data packet;

[0085] The edge server 200 is used to receive and parse the task information and workflow information, and then send a data collection instruction to the edge device 300; process the image data according to the workflow information, and upload the processed image data to the cloud server 100;

[0086] The edge device 300 is used to collect data according to the data collection instruction, obtain image data, and then transmit the image data back to the edge server 200.

[0087] In one embodiment, the cloud server 100 is further used to set task information according to the selection instruction, select a data processing component according to the selection instruction, and generate workflow information based on the data processing component; wherein the task information includes a task execution frequency and a task effective time; the workflow information includes a data processing flow and a data storage path;

[0088] The edge server 200 is also used to parse the task information to obtain the task execution frequency and task effective time, and issue data collection instructions to the edge device 300 according to the task execution frequency and task effective time; parse the workflow information to obtain the data processing flow and data storage path, and perform data management according to the data processing flow and data storage path.

[0089] In one embodiment, the cloud server 100 is also used to receive image data uploaded by the edge server 200, and determine whether a preset receiving mechanism is triggered; wherein the preset receiving mechanism includes any one or more of a data growth rate threshold mechanism, a data cycle mechanism, and a data quantity mechanism; if it is determined that the preset receiving mechanism is not triggered, continue to receive the image data; if it is determined that the preset receiving mechanism is triggered, stop receiving the image data, and send a stop upload instruction to the edge server 200, so that the edge server 200 stops uploading the image data.

[0090] Since the embodiments of the system part correspond to the embodiments of the method part, please refer to the description of the embodiments of the method part for the embodiments of the system part, which will not be repeated here.

[0091] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same and similar parts between the various embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of this application, several improvements and modifications can be made to the present application, and these improvements and modifications also fall within the scope of protection of the claims of this application.

[0092] It should also be noted that, in this specification, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises", "comprising" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the statement "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device including the element.

Claims

1. A method for image device data management based on cloud-edge collaboration, characterized in that: include: In response to the user's selection instruction, the cloud server determines the corresponding edge server and edge device according to the selection instruction; The cloud server configures task information and workflow information for the edge server and edge device, and sends the task information and workflow information to the edge server; The edge server receives and parses the task information and workflow information, and then sends a data collection instruction to the edge device; The cloud server configures task information and workflow information for the edge server and edge device, and sends the task information and workflow information to the edge server, including: The cloud server sets task information according to the selection instruction, selects data processing components according to the selection instruction, and generates workflow information based on the data processing components; wherein the task information includes task execution frequency and task effective time; the workflow information includes data processing flow and data storage path; The edge device collects data according to the data collection instruction to obtain image data, and then transmits the image data back to the edge server; The edge server processes the image data according to the workflow information, and uploads the processed image data to the cloud server; The cloud server receives and manages the image data according to a preset receiving mechanism, and constructs the received image data into an image device data packet; The data processing component includes a data quality detection component; The edge server processes the image data according to the workflow information and uploads the processed image data to the cloud server, including: According to the following formula, the image data is subjected to mosaic detection using the data quality detection component: ； Wherein, r represents the result of mosaic detection, g represents the number of pixels of different gray levels in the image data, and s g represents the number of pixels of the largest first n gray levels, h represents the height of the image data, and w represents the width of the image data; Comparing the mosaic detection result with a preset first threshold requirement; If the result of the mosaic detection meets the preset first threshold requirement, retaining the corresponding image data; If the result of the mosaic detection does not meet the preset first threshold requirement, deleting the corresponding image data; The cloud server receives and manages the image data according to a preset receiving mechanism, and constructs the received image data into an image device data packet, including: The cloud server receives the image data uploaded by the edge server, and determines whether a preset receiving mechanism is triggered; wherein the preset receiving mechanism includes any one or more of a data growth rate threshold mechanism, a data cycle mechanism, and a data quantity mechanism; If it is determined that the preset receiving mechanism is not triggered, continue to receive the image data; If it is determined that the preset receiving mechanism is triggered, the reception of the image data is stopped, and a stop uploading instruction is sent to the edge server, so that the edge server stops uploading the image data.

2. The image device data management method based on cloud-edge collaboration according to claim 1 is characterized in that: The edge server receives and parses the task information and workflow information, and then sends a data collection instruction to the edge device, including: The edge server parses the task information to obtain the task execution frequency and the task effective time, and sends a data collection instruction to the edge device according to the task execution frequency and the task effective time; The edge server parses the workflow information to obtain the data processing flow and data storage path, and performs data management according to the data processing flow and data storage path.

3. The image device data management method based on cloud-edge collaboration according to claim 1 is characterized in that: The edge server processes the image data according to the workflow information and uploads the processed image data to the cloud server, further comprising: Performing Gaussian blur processing on the image data, and respectively calculating the gradient value of the image data before and after the Gaussian blur processing; According to the following formula, the clarity of the image data is detected by using the data quality detection component: ； Where score represents the clarity score, N represents the total number of pixels in the image data, x represents the pixel index in the horizontal direction, y represents the pixel index in the vertical direction, and M represents the pixel index in the horizontal direction. O (x, y) represents the gradient value of the image data before Gaussian blur processing, M b (x, y) represents the gradient value of the image data after Gaussian blur processing; comparing the clarity score with a preset second requirement threshold; If the clarity score meets the preset second threshold requirement, retaining the corresponding image data; If the clarity score does not meet the preset second threshold requirement, the corresponding image data is deleted.

4. The image device data management method based on cloud-edge collaboration according to claim 1 is characterized in that: The data processing component also includes a similarity detection component; The edge server processes the image data according to the workflow information and uploads the processed image data to the cloud server, further comprising: The edge server uses a hash algorithm to calculate a hash mean for the image data; Grouping and aggregating the image data according to the ID of the edge device to obtain multiple image data groups; Based on the similarity detection component, similarity calculation is performed on the image data in each image data group, and deduplication processing is performed on the image data according to the result of the similarity calculation.

5. The image device data management method based on cloud-edge collaboration according to claim 1 is characterized in that: After the cloud server receives and manages the image data according to a preset receiving mechanism and constructs the received image data into an image device data packet, the method further includes: The cloud server uses the large model to perform regional label calibration on the image data in the image device data packet; Use image data with region labels to train the specified model and / or algorithm.

6. An image device data management system based on cloud-edge collaboration, characterized in that: Includes cloud servers, edge servers, and edge devices, including: The cloud server is used to respond to the user's selection instruction, determine the corresponding edge server and edge device according to the selection instruction; configure task information and workflow information for the edge server and edge device, and send the task information and workflow information to the edge server; receive and manage the image data according to a preset receiving mechanism, and construct the received image data into an image device data packet; the cloud server is also used to set task information according to the selection instruction, select a data processing component according to the selection instruction, and generate workflow information based on the data processing component; wherein the task information includes task execution frequency and task effectiveness time; the workflow information includes data processing flow and data storage path; The edge server is used to receive and parse the task information and workflow information, and then send a data collection instruction to the edge device; process the image data according to the workflow information, and upload the processed image data to the cloud server; The edge device is used to collect data according to the data collection instruction to obtain image data, and then transmit the image data back to the edge server; The data processing component includes a data quality detection component; The edge server is further used to perform mosaic detection on the image data using the data quality detection component according to the following formula: ； Wherein, r represents the result of mosaic detection, g represents the number of pixels of different gray levels in the image data, and s g represents the number of pixels of the largest first n gray levels, h represents the height of the image data, and w represents the width of the image data; Comparing the mosaic detection result with a preset first threshold requirement; If the result of the mosaic detection meets the preset first threshold requirement, retaining the corresponding image data; If the result of the mosaic detection does not meet the preset first threshold requirement, deleting the corresponding image data; The cloud server is also used to receive image data uploaded by the edge server and determine whether a preset receiving mechanism is triggered; wherein the preset receiving mechanism includes any one or more of a data growth rate threshold mechanism, a data cycle mechanism and a data quantity mechanism; if it is determined that the preset receiving mechanism is not triggered, continue to receive the image data; if it is determined that the preset receiving mechanism is triggered, stop receiving the image data and send a stop upload instruction to the edge server to cause the edge server to stop uploading the image data.

7. The image device data management system based on cloud-edge collaboration according to claim 6 is characterized in that: The edge server is also used to parse the task information to obtain the task execution frequency and task effective time, and issue data collection instructions to the edge device according to the task execution frequency and task effective time; parse the workflow information to obtain the data processing flow and data storage path, and perform data management according to the data processing flow and data storage path.

Images