Intelligent grouping method, system, terminal and storage medium for surveillance video

Abstract

The present invention discloses a method, system, terminal and storage medium for intelligent grouping of surveillance videos, the method comprising: obtaining the names and group names of all surveillances that need to be processed, obtaining a screenshot of the surveillance content corresponding to each surveillance, selecting some images from all the surveillance content screenshots for manual annotation, and adding scene labels; using a third-party tool to obtain the keywords of the content labels, groups and names of all images, and training a surveillance video classification model in combination with the scene labels to obtain a trained surveillance video classification model; obtaining surveillance videos to be classified that need to be classified, inputting the surveillance videos to be classified into the trained surveillance video classification model, and outputting the classification results of the surveillance videos to be classified. The present invention obtains the keywords of the image names and the image content labels, and finds the relationship between these words and the scene, establishes a classification model based on the relationship, and automatically classifies the newly added surveillance videos after being processed by the classification model.

Description

Technical Field

[0001] The present invention relates to the technical field of video surveillance classification, and in particular to an intelligent grouping method, system, terminal and computer-readable storage medium for surveillance videos. Background Art

[0002] With technological advancements, the variety of communication equipment is increasing. In the field of emergency command, media acquisition equipment is diverse, including drones, portable phones, landlines, and surveillance cameras. These media acquisition devices may be used in different regions and scenarios. Traditional management methods, where a superior manages all surveillance cameras at the subordinate level, are not suitable for all scenarios. For example, querying the surveillance cameras of all intersections or all halls of a company is not feasible.

[0003] Therefore, the existing technology still needs to be improved and developed. Summary of the Invention

[0004] The main purpose of the present invention is to provide an intelligent grouping method, system, terminal and computer-readable storage medium for surveillance videos, aiming to solve the problem in the prior art that when managing surveillance videos, the superior manages all the surveillance of the subordinates. This grouping method is not suitable for all scenarios, resulting in unreasonable classification of surveillance videos.

[0005] To achieve the above object, the present invention provides a method for intelligently grouping surveillance videos, the method comprising the following steps:

[0006] Get the names and group names of all monitoring items that need to be processed, and obtain the corresponding monitoring content screenshots for each monitoring item. Select some images from all monitoring content screenshots for manual annotation and add scene labels.

[0007] Use third-party tools to obtain the content labels, grouping, and name keywords of all images, and train the surveillance video classification model based on the scene labels to obtain a trained surveillance video classification model.

[0008] Obtain a surveillance video to be classified, input the surveillance video to be classified into a trained surveillance video classification model, and output a classification result of the surveillance video to be classified.

[0009] Optionally, the intelligent grouping method for surveillance videos, wherein the steps of obtaining the names and group names of all surveillances to be processed, obtaining a screenshot of the surveillance content corresponding to each surveillance, selecting some images from all the surveillance content screenshots for manual annotation, and adding scene labels, specifically include:

[0010] Get the names and group names of all monitoring items that need to be processed, and get a screenshot of the monitoring content corresponding to each monitoring item. The screenshot of the monitoring content is named after the corresponding monitoring item.

[0011] Select some images from all surveillance content screenshots for manual annotation and add scene labels. The manually annotated images are used as the standard results of model classification.

[0012] Optionally, in the intelligent grouping method of surveillance videos, the process of obtaining a screenshot of the surveillance content corresponding to each surveillance includes: receiving the stream, parsing the I frame and saving it as a picture.

[0013] Optionally, the intelligent grouping method for surveillance videos, wherein the method uses a third-party tool to obtain the content tags, grouping and name keywords of all images, and trains a surveillance video classification model in combination with scene tags to obtain a trained surveillance video classification model, specifically includes:

[0014] Use third-party tools to obtain keywords for content tags, groups, and names of all selected images;

[0015] Organize the scene labels that need to be used, write the correspondence between scene labels and keywords and image content labels, train the surveillance video classification model until it meets the requirements, and obtain a trained surveillance video classification model.

[0016] Optionally, in the intelligent grouping method for surveillance videos, the step of training the surveillance video classification model until it meets the requirements specifically includes:

[0017] After each surveillance video classification model is trained, it is compared with the standard set obtained by manual grouping;

[0018] If the difference with the standard set is greater than the preset requirement, the correspondence between keywords and tags and scene tags will be modified until the classification is reasonable;

[0019] For manually marked unreasonable words, choose to delete and add reasonable words to the vocabulary library.

[0020] Optionally, in the intelligent grouping method for surveillance videos, the vocabulary library includes all keywords and tags.

[0021] Optionally, the intelligent grouping method for surveillance videos, wherein obtaining surveillance videos to be classified, inputting the surveillance videos to be classified into a trained surveillance video classification model, and outputting classification results of the surveillance videos to be classified, specifically includes:

[0022] Obtain a surveillance video to be classified, and input the surveillance video to be classified into a surveillance video classification model with a model accuracy that meets the requirements;

[0023] The surveillance video to be classified is marked with a scene label by a surveillance video classification model, and a classification result of the surveillance video to be classified is output. The classification result is obtained by grouping according to the relationship between the scene label and the image.

[0024] In addition, to achieve the above-mentioned object, the present invention further provides an intelligent grouping system for monitoring videos, wherein the intelligent grouping system for monitoring videos comprises:

[0025] The data processing module is used to obtain the names and group names of all monitoring items that need to be processed, obtain the monitoring content screenshots corresponding to each monitoring item, select some images from all monitoring content screenshots for manual annotation, and add scene labels;

[0026] The model training module is used to obtain the content labels, grouping and name keywords of all images using third-party tools, and train the surveillance video classification model in combination with the scene labels to obtain a trained surveillance video classification model;

[0027] The video classification module is used to obtain a surveillance video to be classified, input the surveillance video to be classified into a trained surveillance video classification model, and output a classification result of the surveillance video to be classified.

[0028] In addition, to achieve the above-mentioned purpose, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and an intelligent grouping program for surveillance videos stored on the memory and runnable on the processor, and when the intelligent grouping program for surveillance videos is executed by the processor, the steps of the intelligent grouping method for surveillance videos as described above are implemented.

[0029] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores an intelligent grouping program for surveillance videos, and when the intelligent grouping program for surveillance videos is executed by a processor, the steps of the intelligent grouping method for surveillance videos as described above are implemented.

[0030] In the present invention, the names and group names of all monitoring that need to be processed are obtained, and the monitoring content screenshots corresponding to each monitoring are obtained. Some images are selected from all monitoring content screenshots for manual annotation and are marked with scene labels; a third-party tool is used to obtain the content labels, grouping and name keywords of all images, and the monitoring video classification model is trained in combination with the scene labels to obtain a trained monitoring video classification model; the monitoring videos to be classified that need to be classified are obtained, the monitoring videos to be classified are input into the trained monitoring video classification model, and the classification results of the monitoring videos to be classified are output. The present invention uses an intelligent method to group monitoring videos, obtain the keywords of the image names and the image content labels, and find the relationship between these words and scenes. Finally, a classification model is established through the relationship. After the new monitoring is added, it will be automatically classified after being processed by the classification model. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a flow chart of a preferred embodiment of the intelligent grouping method for monitoring videos of the present invention;

[0032] Figure 2 It is a flow chart of the data acquisition, model training and monitoring classification process in a preferred embodiment of the intelligent grouping method of monitoring video of the present invention;

[0033] Figure 3 1 is a structural diagram of a preferred embodiment of the intelligent grouping system for monitoring video of the present invention;

[0034] Figure 4 FIG. 4 is a structural diagram of a preferred embodiment of the terminal of the present invention. DETAILED DESCRIPTION

[0035] In order to make the purpose, technical solutions and advantages of the present invention more clear and distinct, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0036] The intelligent grouping method of surveillance video described in the preferred embodiment of the present invention is as follows: Figure 1 and Figure 2 As shown, the intelligent grouping method of surveillance videos includes the following steps:

[0037] Step S10: Obtain the names and group names of all monitoring items that need to be processed, and obtain a screenshot of the monitoring content corresponding to each monitoring item. Select some images from all the monitoring content screenshots for manual annotation and add scene labels.

[0038] Specifically, obtain the names and group names of all monitoring that need to be processed (there are only all groups in advance, and no groups are set for monitoring. The group name is so that AI can recognize image classification. Generally, monitoring is grouped by area. The advantage of this is that it provides another dimension of monitoring grouping for easy viewing), and obtain the monitoring content screenshots corresponding to each monitoring (take one or more pictures, all monitoring in the system has a unique ID, screenshot name: Id_name_serial number.png, so that the group processing results are associated with specific monitoring. The core of obtaining screenshots is completed by ffmpeg, which mainly includes three steps: receiving the stream, parsing the I frame, and saving it as a picture). The monitoring content screenshots are named after the corresponding monitoring; some images are selected from all monitoring content screenshots for manual annotation, and scene labels are added, and the manually annotated images are used as the standard results of model classification.

[0039] Step S20: Use a third-party tool to obtain the content tags, grouping and name keywords of all images, and train the surveillance video classification model in combination with the scene tags to obtain a trained surveillance video classification model.

[0040] Specifically, a third-party tool is used to obtain the content labels, groups, and keywords of the names of all selected images. Keywords are the result output of automatic recognition of images by AI. Untrained keywords may not be used as groups, and keywords with the highest scores after training can be used as groups. By decomposing phrases in monitoring names, groupings required by the emergency industry may also be obtained. For example, names containing elementary school, school, etc. For example, keywords include: cabinet, indoor corner, display screen, carpet, brick and stone, bag, thick fog, astronomical photography, polished tile, aluminum-plastic panel, birds, elevator, modern architecture, machinery and equipment, bicycle, canyon, etc. During model training, manual labels are added. The second recognition result will include manual labels and some picture vocabulary. If the manual label is still not ranked first, more image content will be annotated from the picture and retraining will be performed until the first or first few recognition results are the expected content labels. The scene labels needed are sorted out, and the correspondence between scene labels and keywords and image content labels is compiled. The surveillance video classification model is trained until it meets the requirements to obtain a trained surveillance video classification model.

[0041] Iterate according to the obtained keywords and image content labels until a suitable model is obtained; organize commonly used scene labels, image content labels and keywords, and words with low probability of appearing cannot be used as training data sets. According to the keywords and image content labels of the manually annotated images, manually write rules to train the model (for example: ponds belong to agriculture, forestry, animal husbandry and fishery, intersections belong to transportation, gas stations and tank areas belong to hazardous chemicals). Test the accuracy of the model until it meets the requirements. After each model training is completed, compare it with the standard set obtained by manual grouping (that is, the correspondence between the monitored corresponding images and scene labels); if there is a significant gap with the standard set, modify the correspondence between keywords and labels and scene labels until the classification is basically reasonable; for unreasonable manually annotated words, choose to delete them. Similarly, reasonable words are added to the vocabulary library (including all keywords and labels).

[0042] This is the model training process. Basic AI recognizes individual items, such as spoons, stoves, tables, signs, pipes, oil guns, and fire extinguishers. However, the AI needs to be able to recognize kitchens, schools, gas stations, and the like. It will find many images of different kitchens, annotate the important distinguishing elements in the images, and assign kitchen groups. These images are then trained on third-party tools, ensuring that the groups they recognize include the given labels. Repeated training increases the scores of the defined labels. After training, the third-party tools will recognize similar images and assign the desired groupings. The training images are manually selected, and both the image identification and labeling are done manually, with images of one category selected at a time. Manual scene labeling is a key classification practice in the emergency response industry.

[0043] Choose appropriate development tools; use a standard 4-core 16GB memory server with a Windows operating system to develop the software, use VScode to compile the software, and use third-party model training tools (such as Baidu Feipang), third-party libraries (such as ffmpeg), and image annotation tools.

[0044] Step S30: Obtain a surveillance video to be classified, input the surveillance video to be classified into a trained surveillance video classification model, and output a classification result of the surveillance video to be classified.

[0045] Specifically, the system obtains surveillance videos to be classified and feeds them into a surveillance video classification model that meets the required accuracy. The model then assigns scene labels to the videos and outputs classification results. The classification results are grouped based on the relationship between the scene labels and the images. Furthermore, manual grouping is used for surveillance videos that cannot be grouped.

[0046] The technical effects that the present invention can bring are as follows:

[0047] (1) Efficiency: After the model is established, when new monitoring is added to the management, it will be automatically classified according to the rules set in the code.

[0048] (2) Plasticity: When there are new requirements, you can add a scene recognition by modifying the model. Similarly, if you are not satisfied with the existing classification model, you can directly modify the model.

[0049] (3) Easy to iterate: The model is mainly used for classification, without excessive calculations and deductions, so the algorithm will not be significantly modified. During the iteration process, a new model can be directly obtained by adjusting the dataset of the training model, and then a new grouping can be obtained.

[0050] (4) Applicable to multiple scenarios: Since the classification rules are determined by the model, as long as there is a data set, it can theoretically be applied to various scenarios.

[0051] The present invention adopts an intelligent way to perform grouping. First, the keywords of the image name and the image content label are obtained, and then the relationship between these words and scenes (such as agriculture, animal husbandry and fishery, medical care and health) is found. Finally, a model is established through the relationship. In actual application, after the new monitoring is added, it will be automatically classified after being processed by the model.

[0052] Further, if Figure 3 As shown, based on the above-mentioned intelligent grouping method of surveillance videos, the present invention also provides an intelligent grouping system for surveillance videos, wherein the intelligent grouping system for surveillance videos includes:

[0053] The data processing module 51 is used to obtain the names and group names of all monitoring items to be processed, obtain the monitoring content screenshots corresponding to each monitoring item, select some images from all monitoring content screenshots for manual annotation, and add scene labels;

[0054] The model training module 52 is used to obtain the content labels, grouping and name keywords of all images using a third-party tool, and train the surveillance video classification model in combination with the scene labels to obtain a trained surveillance video classification model;

[0055] The video classification module 53 is used to obtain a surveillance video to be classified, input the surveillance video to be classified into a trained surveillance video classification model, and output a classification result of the surveillance video to be classified.

[0056] Further, if Figure 4 As shown, based on the above-mentioned intelligent grouping method and system for surveillance videos, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 4Only some of the components of the terminal are shown, but it should be understood that implementation of all of the shown components is not required, and more or fewer components may be implemented instead.

[0057] In some embodiments, the memory 20 can be an internal storage unit of the terminal, such as a hard disk or memory of the terminal. In other embodiments, the memory 20 can also be an external storage device of the terminal, such as a plug-in hard disk equipped on the terminal, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. Furthermore, the memory 20 can also include both an internal storage unit of the terminal and an external storage device. The memory 20 is used to store application software and various types of data installed on the terminal, such as the program code of the installation terminal. The memory 20 can also be used to temporarily store data that has been output or is to be output. In one embodiment, the memory 20 stores an intelligent grouping program 40 for monitoring video, and the intelligent grouping program 40 for monitoring video can be executed by the processor 10, thereby realizing the intelligent grouping method for monitoring video in the present application.

[0058] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, configured to execute program codes or process data stored in the memory 20, such as executing the intelligent grouping method for surveillance videos.

[0059] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface. The components 10-30 of the terminal communicate with each other via a system bus.

[0060] In one embodiment, when the processor 10 executes the intelligent grouping program 40 of the surveillance videos in the memory 20, the following steps are implemented:

[0061] Get the names and group names of all monitoring items that need to be processed, and obtain the corresponding monitoring content screenshots for each monitoring item. Select some images from all monitoring content screenshots for manual annotation and add scene labels.

[0062] Use third-party tools to obtain the content labels, grouping, and name keywords of all images, and train the surveillance video classification model based on the scene labels to obtain a trained surveillance video classification model.

[0063] Obtain a surveillance video to be classified, input the surveillance video to be classified into a trained surveillance video classification model, and output a classification result of the surveillance video to be classified.

[0064] The process of obtaining the names and group names of all monitoring items to be processed, obtaining screenshots of monitoring content corresponding to each monitoring item, selecting some images from all monitoring content screenshots for manual annotation, and adding scene labels specifically includes:

[0065] Get the names and group names of all monitoring items that need to be processed, and get a screenshot of the monitoring content corresponding to each monitoring item. The screenshot of the monitoring content is named after the corresponding monitoring item.

[0066] Select some images from all surveillance content screenshots for manual annotation and add scene labels. The manually annotated images are used as the standard results of model classification.

[0067] The process of obtaining a screenshot of the monitoring content corresponding to each monitoring includes: receiving the stream, parsing the I frame and saving it as a picture.

[0068] The method of using a third-party tool to obtain the content tags, grouping and name keywords of all images, and training the surveillance video classification model in combination with the scene tags to obtain a trained surveillance video classification model specifically includes:

[0069] Use third-party tools to obtain keywords for content tags, groups, and names of all selected images;

[0070] Organize the scene labels that need to be used, write the correspondence between scene labels and keywords and image content labels, train the surveillance video classification model until it meets the requirements, and obtain a trained surveillance video classification model.

[0071] The training of the surveillance video classification model until it meets the requirements specifically includes:

[0072] After each surveillance video classification model is trained, it is compared with the standard set obtained by manual grouping;

[0073] If the difference with the standard set is greater than the preset requirement, the correspondence between keywords and tags and scene tags will be modified until the classification is reasonable;

[0074] For manually marked unreasonable words, choose to delete and add reasonable words to the vocabulary library.

[0075] The vocabulary library includes all keywords and tags.

[0076] The step of obtaining a surveillance video to be classified, inputting the surveillance video to be classified into a trained surveillance video classification model, and outputting a classification result of the surveillance video to be classified specifically includes:

[0077] Obtain a surveillance video to be classified, and input the surveillance video to be classified into a surveillance video classification model with a model accuracy that meets the requirements;

[0078] The surveillance video to be classified is marked with a scene label by a surveillance video classification model, and a classification result of the surveillance video to be classified is output. The classification result is obtained by grouping according to the relationship between the scene label and the image.

[0079] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores an intelligent grouping program for surveillance videos, and when the intelligent grouping program for surveillance videos is executed by a processor, the steps of the intelligent grouping method for surveillance videos as described above are implemented.

[0080] In summary, the present invention provides a method, system, terminal and storage medium for intelligent grouping of surveillance videos, the method comprising: obtaining the names and group names of all surveillances that need to be processed, and obtaining a screenshot of the surveillance content corresponding to each surveillance, selecting some images from all the surveillance content screenshots for manual annotation, and marking them with scene labels; using a third-party tool to obtain the content labels, grouping and name keywords of all images, and training a surveillance video classification model in combination with the scene labels to obtain a trained surveillance video classification model; obtaining surveillance videos to be classified that need to be classified, inputting the surveillance videos to be classified into the trained surveillance video classification model, and outputting the classification results of the surveillance videos to be classified. The present invention uses an intelligent method to group surveillance videos, obtains keywords of image names and image content labels, and finds the relationship between these words and scenes, and finally establishes a classification model through the relationship. After adding surveillance, it will be automatically classified after being processed by the classification model.

[0081] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or terminal comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or terminal. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or terminal comprising the element.

[0082] Of course, those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware (such as a processor, controller, etc.) through a computer program. The program can be stored in a computer-readable storage medium that can be read by a computer. When the program is executed, it can include the processes in the above-described method embodiments. The computer-readable storage medium can be a memory, a magnetic disk, an optical disk, etc.

[0083] It should be understood that the application of the present invention is not limited to the above examples. For those skilled in the art, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A method for intelligent grouping of surveillance videos, characterized in that: The intelligent grouping method of the surveillance video includes: Get the names and group names of all monitoring items that need to be processed, and obtain the corresponding monitoring content screenshots for each monitoring item. Select some images from all monitoring content screenshots for manual annotation and add scene labels. Use third-party tools to obtain the content labels, grouping, and name keywords of all images, and train the surveillance video classification model based on the scene labels to obtain a trained surveillance video classification model. Obtain a surveillance video to be classified, input the surveillance video to be classified into a trained surveillance video classification model, and output a classification result of the surveillance video to be classified; The method of using a third-party tool to obtain the content tags, grouping and name keywords of all images and training the surveillance video classification model in combination with the scene tags to obtain a trained surveillance video classification model specifically includes: Use third-party tools to obtain keywords for content tags, groups, and names of all selected images; Organize the scene labels needed, compile the correspondence between scene labels, keywords, and image content labels, train the surveillance video classification model until it meets the requirements, and obtain a trained surveillance video classification model; When training the surveillance video classification model, adding manual labels; The monitoring video classification model is trained until it meets the requirements, specifically including: After each surveillance video classification model is trained, it is compared with the standard set obtained by manual grouping; If the difference with the standard set is greater than the preset requirement, the correspondence between keywords and tags and scene tags will be modified until the classification is reasonable; For manually marked unreasonable words, choose to delete them and add reasonable words to the vocabulary library; An intelligent approach is used to group images, obtain keywords from image names and image content tags, find the relationship between vocabulary and scenes, and build a model based on the relationship. In practical applications, when new monitoring is added, it will be automatically classified after being processed by the model.

2. The intelligent grouping method for surveillance videos according to claim 1, characterized in that: The process of obtaining the names and group names of all monitoring items that need to be processed, obtaining screenshots of the monitoring content corresponding to each monitoring item, selecting some images from all monitoring content screenshots for manual annotation, and adding scene labels specifically includes: Get the names and group names of all monitoring items that need to be processed, and get a screenshot of the monitoring content corresponding to each monitoring item. The screenshot of the monitoring content is named after the corresponding monitoring item. Select some images from all surveillance content screenshots for manual annotation and add scene labels. The manually annotated images are used as the standard results of model classification.

3. The intelligent grouping method for surveillance videos according to claim 2, characterized in that: The process of obtaining a screenshot of the monitoring content corresponding to each monitoring includes: receiving the stream, parsing the I frame and saving it as a picture.

4. The intelligent grouping method for surveillance videos according to claim 1, characterized in that: The vocabulary includes all keywords and tags.

5. The intelligent grouping method for surveillance videos according to claim 1, characterized in that: The step of obtaining a surveillance video to be classified, inputting the surveillance video to be classified into a trained surveillance video classification model, and outputting a classification result of the surveillance video to be classified specifically includes: Obtain a surveillance video to be classified, and input the surveillance video to be classified into a surveillance video classification model with a model accuracy that meets the requirements; The surveillance video to be classified is marked with a scene label by a surveillance video classification model, and a classification result of the surveillance video to be classified is output. The classification result is obtained by grouping according to the relationship between the scene label and the image.

6. An intelligent grouping system for surveillance video, characterized in that: The intelligent grouping system for surveillance videos is applied to the intelligent grouping method for surveillance videos according to any one of claims 1 to 5, and the intelligent grouping system for surveillance videos includes: The data processing module is used to obtain the names and group names of all monitoring items that need to be processed, obtain the monitoring content screenshots corresponding to each monitoring item, select some images from all monitoring content screenshots for manual annotation, and add scene labels; The model training module is used to obtain the content labels, grouping and name keywords of all images using third-party tools, and train the surveillance video classification model in combination with the scene labels to obtain a trained surveillance video classification model; The video classification module is used to obtain a surveillance video to be classified, input the surveillance video to be classified into a trained surveillance video classification model, and output a classification result of the surveillance video to be classified.

7. A terminal, characterized in that: The terminal includes: a memory, a processor, and an intelligent grouping program for surveillance videos stored in the memory and runnable on the processor. When the intelligent grouping program for surveillance videos is executed by the processor, the steps of the intelligent grouping method for surveillance videos as described in any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program for intelligent grouping of surveillance videos, and when the program for intelligent grouping of surveillance videos is executed by a processor, the steps of the method for intelligent grouping of surveillance videos as described in any one of claims 1 to 5 are implemented.

Images