A method of image archiving and related apparatus

Abstract

Embodiments of the present application disclose a kind of image filing method and related device, it is characterized in that, the method comprises: obtaining the image information based on M camera;According to image information, determine N travel trajectory;N travel trajectory is associated with L camera set;Each camera set in L camera set includes one or more travel trajectory in N travel trajectory in M camera covered camera;According to L camera set, generate L high-frequency archive database;Each high-frequency archive database includes multiple high-frequency archives;Each high-frequency archive corresponds to the image set of one high-frequency person, each high-frequency person is the person whose frequency of appearance in each high-frequency archive database matched camera set exceeds preset value or is the person whose frequency of appearance is in front;L high-frequency archive database is used to file the image collected by target camera. Using the embodiment of the present application can improve image filing efficiency, improve user experience.

Description

Technical Field

[0001] This invention relates to the field of information processing, and more particularly to a method and related apparatus for image archiving. Background Technology

[0002] With the rapid development of facial recognition technology, its application scenarios are becoming increasingly widespread. In urban video surveillance, a large number of facial images can be extracted from the video, and images similar to a given face can be found from a massive database of facial images. However, this single retrieval method requires a large amount of similarity calculation, placing a significant computational burden on the system. Furthermore, this method often necessitates manual review and analysis to obtain more relevant information; therefore, it is only suitable for a limited number of applications.

[0003] Currently, in existing technologies, facial images can be organized through real-time archiving. These archives allow for efficient acquisition of human activity trajectories, enabling the extraction of more valuable information. The real-time archiving method involves comparing the features of each captured facial image with existing facial image archives in real time to determine the archive to which the image belongs, thus achieving real-time archiving of facial images. However, in large-scale real-time archiving scenarios, this method is limited by hardware processing and computing power, resulting in long archiving times and low efficiency.

[0004] Therefore, improving image archiving efficiency is an urgent problem to be solved. Summary of the Invention

[0005] The technical problem to be solved by the embodiments of the present invention is to provide an image archiving method and related apparatus, which can improve image archiving efficiency and enhance user experience.

[0006] In a first aspect, embodiments of the present invention provide an image archiving method, characterized in that the method includes: acquiring image information based on M cameras; determining N travel trajectories based on the image information; the N travel trajectories being associated with L camera sets; wherein each of the L camera sets includes one or more travel trajectories from the N travel trajectories covering cameras in the M cameras; M is an integer greater than 0; N is an integer greater than 0; L is an integer greater than 0; generating L high-frequency archive bases based on the L camera sets; wherein one high-frequency archive base matches one camera set; each high-frequency archive base includes multiple high-frequency archives; each high-frequency archive corresponds to an image set of a high-frequency person, and each high-frequency person is a person whose frequency of appearance in the camera set matched by each high-frequency archive base exceeds a preset value or whose appearance frequency is among the highest; the L high-frequency archive bases are used to archive images captured by a target camera, the target camera being any one of the M cameras.

[0007] In large-scale real-time image archiving scenarios, existing technologies can employ a hierarchical archiving method. This involves first comparing and archiving images against a high-frequency archive database. If no similar images are found in the high-frequency archive database, the images are then compared and archived against the full archive database. However, existing technologies only group multiple cameras based on region, and for each camera group, use its corresponding regional resident population archive database as the high-frequency archive database. For example, all cameras in Pudong New Area, Shanghai, are grouped together, and the Pudong New Area resident population archive database is used as the high-frequency archive database. Then, based on this high-frequency archive database, real-time image archiving is performed on facial images captured by cameras covering Pudong New Area. If archiving fails, the Shanghai resident population archive database is used as the full archive database, and real-time image archiving is then performed based on this full archive database. Because the regional permanent resident population database is updated infrequently and does not record the files of the floating population, in areas with high population mobility, a large number of facial images cannot be archived at the stage of comparison with the high-frequency archive database. Instead, they need to be compared with the full archive database before being archived. This results in a large amount of computation and low efficiency in the image archiving process.

[0008] In this embodiment of the invention, pedestrian trajectory information can be determined based on the image information obtained by the camera. Further, based on this pedestrian trajectory information, the space is divided into multiple partitions, and cameras covered by the same partition are grouped together, thereby enabling adjacent cameras to be assigned to the same group. Then, based on the camera grouping information and image information, high-frequency files appearing in each partition can be counted to obtain a high-frequency file base. Further, image archiving can be performed based on this high-frequency file base, for example, hierarchical image archiving, that is, first comparing the images captured by the camera with the files in the high-frequency file base. If no similar files are found, then comparing them with the files in the full file base and archiving them. In summary, unlike existing technologies that group cameras solely based on region and directly use the regional resident population database as the high-frequency archive base, this invention divides the space based on pedestrian trajectory information, grouping cameras covered by the same partition together. It also statistically analyzes high-frequency archives for different partitions. This not only allows for the periodic maintenance of camera groups in each partition but also enables the periodic updating of the high-frequency archive base for each partition. This ensures that the high-frequency archive base includes population records that have frequently appeared in the region recently. Consequently, a large number of facial images can be directly archived at the first level of hierarchical image archiving (such as the comparison stage with the high-frequency archive base), effectively reducing computational load, improving image archiving efficiency, and enhancing user experience. Furthermore, pre-archiving using the regional high-frequency archive base also increases the probability of images being correctly assigned to the archive and improves computational accuracy.

[0009] In one possible implementation, the image information includes archive file identifiers of multiple archived images, camera identifier information of the multiple archived images, and archive time information of the multiple archived images; determining N travel trajectories based on the image information includes: determining the N travel trajectories based on the camera identifier information and the archive time information corresponding to each archive file identifier.

[0010] In this embodiment of the invention, when the image information includes archive file identifiers, camera identifier information, and archive time information of multiple archived images, images with the same archive file identifier are processed based on the camera identifier information and archive time information to obtain multiple pedestrian trajectories. Then, cameras with high trajectory repetition can be grouped together based on these pedestrian trajectories. By determining pedestrian trajectories based on relevant information of archived images using the method provided in this embodiment of the invention, pedestrian trajectories that are closer to the actual situation can be obtained, thereby making camera grouping more reasonable, improving image archiving efficiency, and enhancing user experience.

[0011] In one possible implementation, the method further includes: determining the L sets of cameras based on the N travel trajectories.

[0012] In this embodiment of the invention, after confirming multiple pedestrian trajectories, multiple partitions can be obtained by spatial division based on these pedestrian trajectories. Then, cameras covered by the same partition are grouped together to obtain multiple camera sets. The method provided by this embodiment of the invention, by grouping multiple cameras, decomposes large-scale real-time image archiving into multiple small-scale real-time image archiving that can be performed simultaneously and in parallel, thereby improving computational efficiency, improving image archiving efficiency, and enhancing user experience.

[0013] In one possible implementation, the image information includes camera identification information for multiple archived images; determining the L camera set based on the N travel trajectories includes: using the camera identification information of the multiple archived images to calculate the data processing volume of each of the M cameras within a first preset time period; and determining the L camera set based on the data processing volume of each camera and the N travel trajectories.

[0014] In this embodiment of the invention, since the image information includes camera identification information for multiple archived images, the data processing volume of each camera within a preset time period (e.g., the data processing volume over the past seven days) can be calculated based on this camera identification information. Then, based on the data processing volume of each camera within the preset time period and the obtained pedestrian trajectories, the multiple cameras are grouped to obtain multiple camera sets. Because both pedestrian trajectories and the data processing volume of each camera within the preset time period are considered when grouping the cameras, it is possible to group cameras covered by densely populated areas together, with the boundaries between groups located in sparsely populated areas. This avoids excessive duplication of high-frequency archive databases in different areas, improves image archiving efficiency, and enhances the user experience.

[0015] In one possible implementation, the method further includes: acquiring a first image of a target object using the target camera; determining a target camera set corresponding to the target camera from the L camera sets, and determining a target high-frequency archive database corresponding to the target camera set from the L high-frequency archive databases; determining whether there is an archive similar to the first image in the target high-frequency archive database; if so, archiving the first image into the archive, and recording the archived archive identifier, camera identifier information, and archived time information of the first image.

[0016] In this embodiment of the invention, when a target camera captures a facial image, it can first determine whether a similar file exists in the high-frequency archive database corresponding to the camera set to which the target camera belongs. If so, the captured facial image is archived into that file. Simultaneously, information such as the archived file identifier, the target camera identifier, and the archive time can be recorded, facilitating subsequent updates to the camera set and its corresponding high-frequency archive database. By comparing and archiving the facial image captured by the target camera with the aforementioned high-frequency archive database, the captured facial image is likely archived during the comparison stage with the high-frequency archive database, reducing the need for comparison with the full archive database, lowering computational load, improving image archiving efficiency, and enhancing user experience.

[0017] In one possible implementation, the method further includes: if there is no file similar to the first image in the target high-frequency archive database; then determine whether there is a file similar to the first image in the full archive database; the full archive database includes files of people who have appeared under the M cameras; if there is, then archive the first image into the file, and record the archived file identifier, camera identifier information and archived time information of the first image.

[0018] In this embodiment of the invention, when there are no similar files in the high-frequency archive database corresponding to the camera set of the target camera, the captured facial image can be compared with the full archive database (such as the permanent resident population database and the floating population database of the entire city, or all historical files that have appeared under all cameras). If there are similar files in the full archive database, the captured facial image is assigned to that file. At the same time, information such as the archived file identifier, the target camera identifier, and the archived time can be recorded, which can facilitate subsequent updates to the camera set and its corresponding high-frequency archive database, improve image archiving efficiency, and enhance user experience.

[0019] In one possible implementation, the method further includes: if there is no file similar to the first image in the full archive database; then create a new file, archive the first image into the new file, and record the archive file identifier, camera identifier information and archive time information of the first image.

[0020] In this embodiment of the invention, when no similar file exists in the full archive database, a new file can be created, and the image can be archived into the newly created file. At the same time, information such as the archived file identifier, the target camera identifier, and the archived time can be recorded, which facilitates subsequent updates to the camera set and its corresponding high-frequency archive database, improves image archiving efficiency, and enhances user experience.

[0021] In one possible implementation, the method further includes: obtaining a target full archive base based on the archived files, determining whether similar files exist in the target full archive base; if so, merging the similar files.

[0022] In this embodiment of the invention, after the images captured by the camera are archived, a target full archive database can be obtained based on the archived files, and it can be determined whether there are similar files in the target full archive database. If there are, the similar files can be merged, thereby reducing the problem of file duplication, improving image archiving efficiency, and enhancing user experience.

[0023] In one possible implementation, the method further includes: deleting files in the target full archive that have not been updated within a second preset time period.

[0024] In this embodiment of the invention, if there are files in the target full archive database that have not been updated for a long time, these files can be deleted from the full archive database. This not only saves storage resources, but also reasonably reduces the number of files contained in the full archive database, thereby reducing unnecessary computational overhead when archiving images, thus improving image archiving efficiency and enhancing user experience.

[0025] Secondly, embodiments of the present invention provide an image archiving method, characterized in that the method includes: a main server acquiring image information based on M cameras; the main server determining N travel trajectories based on the image information; the N travel trajectories being associated with L camera sets; wherein each of the L camera sets includes one or more travel trajectories from the N travel trajectories that are covered by the M cameras; M is an integer greater than 0; N is an integer greater than 0; and L is an integer greater than 0; the main server generating L high-frequency archive bases based on the L camera sets; wherein one high-frequency archive base matches one camera set; each high-frequency archive base includes multiple high-frequency archives; and each high-frequency archive corresponds to a high-frequency person. The image set, where each high-frequency person is a person whose frequency of appearance exceeds a preset value or whose appearance frequency is among the highest in each high-frequency archive database matched by the camera set; the sub-server acquires a first image of the target object through the target camera; the target camera is any one of the M cameras; the sub-server determines the target camera set corresponding to the target camera from the L camera sets, and determines the target high-frequency archive database corresponding to the target camera set from the L high-frequency archive databases; the sub-server determines whether there is an archive similar to the first image in the target high-frequency archive database; if so, the sub-server archives the first image into the archive and records the archived archive identifier, camera identifier information and archived time information of the first image.

[0026] In this embodiment of the invention, the main server determines pedestrian trajectory information based on image information, further divides the space into multiple partitions based on this pedestrian trajectory information, and groups cameras covered by the same partition together, thereby assigning adjacent cameras to the same group. Then, based on the camera grouping information and image information, high-frequency files appearing in each partition can be counted to obtain a high-frequency file database. Sub-servers can then perform hierarchical image archiving based on this high-frequency file database. In summary, unlike existing technologies that only group cameras by region and directly use the regional resident population database as the high-frequency file database, this embodiment of the invention divides the space based on pedestrian trajectory information, groups cameras covered by the same partition together, and counts regional high-frequency files for different partitions. This not only allows for regular maintenance of camera groups in each partition but also regular updates to the high-frequency file database for each partition, ensuring that the high-frequency file database contains population files that frequently appear in the region. This allows sub-servers to directly archive a large number of face images at the first level of hierarchical image archiving (such as the comparison stage with the high-frequency file database), effectively reducing computational load, improving image archiving efficiency, and enhancing user experience. Meanwhile, the main server is responsible for grouping cameras and compiling a high-frequency archive database, while the sub-servers are responsible for archiving images in real time. This allows multiple sub-servers to archive images simultaneously, achieving distributed computing and improving image archiving efficiency.

[0027] Thirdly, embodiments of the present invention provide an image archiving apparatus, characterized in that the apparatus comprises: a first acquisition unit, configured to acquire image information obtained from M cameras; a first processing unit, configured to determine N travel trajectories based on the image information; the N travel trajectories are associated with L camera sets; wherein each of the L camera sets includes one or more travel trajectories from the N travel trajectories that cover the cameras in the M cameras; M is an integer greater than 0; N is an integer greater than 0; and L is an integer greater than 0; a second processing unit, configured to generate L high-frequency archive bases based on the L camera sets; wherein one high-frequency archive base matches one camera set; each high-frequency archive base includes multiple high-frequency archives; each high-frequency archive corresponds to an image set of a high-frequency person, and each high-frequency person is a person whose frequency of appearance in the camera set matched by each high-frequency archive base exceeds a preset value or whose appearance frequency is among the highest; the L high-frequency archive bases are used to archive images captured by a target camera, the target camera being any one of the M cameras.

[0028] In one possible implementation, the image information includes archive file identifiers of multiple archived images, camera identifier information of the multiple archived images, and archive time information of the multiple archived images; the first processing unit is specifically used to: determine the N travel trajectories based on the camera identifier information and the archive time information corresponding to each of the archive file identifiers.

[0029] In one possible implementation, the device further includes a third processing unit for determining the L sets of cameras based on the N travel trajectories.

[0030] In one possible implementation, the image information includes camera identification information of multiple archived images; the third processing unit is specifically used to: use the camera identification information of the multiple archived images to count the data processing volume of each of the M cameras within a first preset time period; and determine the L camera set based on the data processing volume of each camera and the N travel trajectories.

[0031] In one possible implementation, the apparatus further includes: a second acquisition unit, configured to acquire a first image of a target object using the target camera; a fourth processing unit, configured to determine a target camera set corresponding to the target camera from the L camera sets, and to determine a target high-frequency archive database corresponding to the target camera set from the L high-frequency archive databases; a fifth processing unit, configured to determine whether there is an archive similar to the first image in the target high-frequency archive database; and a first archiving unit, configured to archive the first image into the archive if it exists, and record the archived archive identifier, camera identifier information, and archiving time information of the first image.

[0032] In one possible implementation, the device further includes: a second archiving unit, configured to determine whether a file similar to the first image exists in the full archive if no file similar to the first image exists in the target high-frequency archive database; the full archive database includes files of people who have appeared under the M cameras; if such a file exists, the first image is archived in the archive, and the archived file identifier, camera identifier information, and archived time information of the first image are recorded.

[0033] In one possible implementation, the device further includes: a third archiving unit, configured to create a new archive if no archive similar to the first image exists in the full archive database, archive the first image into the new archive, and record the archive identifier, camera identifier information, and archive time information of the first image.

[0034] In one possible implementation, the apparatus further includes: a first merging unit, configured to obtain a target full archive database based on the archived files, determine whether similar files exist in the target full archive database; if so, merge the similar files.

[0035] In one possible implementation, the apparatus further includes: a first deletion unit, used to delete files in the target full-volume archive that have not been updated within a second preset time period.

[0036] Fourthly, embodiments of the present invention provide an image archiving apparatus, characterized in that the apparatus comprises: a first acquisition unit of a main server, configured to acquire image information obtained from M cameras; a first processing unit of the main server, configured to determine N travel trajectories based on the image information; the N travel trajectories are associated with L camera sets; wherein each of the L camera sets includes one or more cameras covered by the N travel trajectories in the M cameras; M is an integer greater than 0; N is an integer greater than 0; and L is an integer greater than 0; a second processing unit of the main server, configured to generate L high-frequency archive bases based on the L camera sets; wherein one high-frequency archive base matches one camera set; each high-frequency archive base includes multiple high-frequency archives; and each high-frequency archive corresponds to a high-frequency person. The image set of the sub-server, wherein each high-frequency person is a person whose frequency of appearance exceeds a preset value or whose appearance frequency is among the highest in the camera set matched in each high-frequency archive database; the second acquisition unit of the sub-server is used to acquire a first image of the target object through the target camera; the target camera is any one of the M cameras; the third processing unit of the sub-server is used to determine the target camera set corresponding to the target camera from the L camera sets, and to determine the target high-frequency archive database corresponding to the target camera set from the L high-frequency archive databases; the first archiving unit of the sub-server is used to determine whether there is an archive similar to the first image in the target high-frequency archive database; if so, the sub-server archives the first image into the archive and records the archived archive identifier, camera identifier information and archived time information of the first image.

[0037] Fifthly, embodiments of the present invention provide a chip system, characterized in that the chip system includes at least one processor, a memory, and an interface circuit, the memory, the interface circuit, and the at least one processor being interconnected via circuits, and the at least one memory storing instructions; when the instructions are executed by the processor, any one of the methods of the first aspect described above is implemented.

[0038] In a sixth aspect, embodiments of the present invention provide a chip system, characterized in that the chip system includes at least one processor, a memory, and an interface circuit, the memory, the interface circuit, and the at least one processor being interconnected via circuits, and the at least one memory storing instructions; when the instructions are executed by the processor, any one of the methods in the second aspect described above is implemented.

[0039] In a seventh aspect, embodiments of the present invention provide a computer storage medium, characterized in that the computer storage medium stores a computer program, which, when executed by a processor, implements the method described in any one of the first aspects above.

[0040] Eighthly, embodiments of the present invention provide a computer storage medium, characterized in that the computer storage medium stores a computer program, which, when executed by a processor, implements the method described in any one of the second aspects above.

[0041] In a ninth aspect, embodiments of the present invention provide a computer program, characterized in that the computer program includes instructions that, when executed by a computer, cause the computer to perform the method described in any one of the first aspects above.

[0042] In a tenth aspect, embodiments of the present invention provide a computer program, characterized in that the computer program includes instructions that, when executed by a computer, cause the computer to perform the method described in any one of the second aspects above.

[0043] Eleventhly, this application provides a terminal device that has the function of implementing any of the image archiving methods provided in the first aspect above. This function can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described function.

[0044] In a twelfth aspect, this application provides a terminal device that has the function of implementing any of the image archiving methods provided in the first aspect above. This function can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described function.

[0045] In a thirteenth aspect, this application provides an intelligent device that has the function of implementing any of the image archiving methods provided in the first aspect above. This function can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described function.

[0046] In a fourteenth aspect, this application provides an intelligent device that has the function of implementing any of the image archiving methods provided in the first aspect above. This function can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described function. Attached Figure Description

[0047] Figure 1A A schematic diagram of a basic real-time archiving process provided for existing technologies.

[0048] Figure 1B A schematic diagram of a hierarchical archiving process provided for existing technologies.

[0049] Figure 1C This is a schematic diagram of an image archiving system architecture provided by an embodiment of the present invention.

[0050] Figure 1D This is a schematic diagram of another image archiving system architecture provided by an embodiment of the present invention.

[0051] Figure 2A This is a flowchart illustrating an image archiving method according to an embodiment of this application.

[0052] Figure 2B This is a flowchart illustrating a process for determining a travel trajectory, as provided in an embodiment of the present invention.

[0053] Figure 2C This is a schematic diagram of camera grouping provided in an embodiment of the present invention.

[0054] Figure 2D This is a schematic diagram of a camera grouping method provided in an embodiment of the present invention.

[0055] Figure 3A This is a flowchart illustrating the real-time image archiving stage of an image archiving method according to an embodiment of this application.

[0056] Figure 3B This is a flowchart illustrating another image archiving method in an embodiment of this application.

[0057] Figure 4A This is an exemplary flowchart of an image archiving method provided in an embodiment of the present invention.

[0058] Figure 4B An exemplary flowchart of another image archiving method provided in an embodiment of the present invention.

[0059] Figure 5A This application provides a schematic diagram of an image archiving device according to an embodiment of the present invention.

[0060] Figure 5BThis application provides a schematic diagram of an image archiving device according to an embodiment of the present invention. Detailed Implementation

[0061] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0062] First, the specific technical problem to be solved in this application is analyzed and proposed. Existing technologies for real-time image archiving include Scheme 1 and Scheme 2. Among them,

[0063] Option 1: Basic real-time archiving method.

[0064] Please see Figure 1A , Figure 1A This diagram illustrates a basic real-time archiving process for existing technologies. After a picture is taken, a facial image is extracted and then compared with the comprehensive features of existing archives (such as the combined result of multiple facial images in an archive) (e.g., similarity calculation). If the similarity is higher than a threshold, the comparison is successful, the image is archived in that archive, and the comprehensive features of that archive are updated. If the similarity is lower than the threshold, the comparison fails, a new archive is created, and the new archive is added to the existing archive.

[0065] Solution 1 can achieve real-time archiving of facial images, but it has the following drawbacks:

[0066] Disadvantage 1: This technology only completes basic real-time archiving. In large-scale real-time archiving scenarios, relying solely on basic real-time archiving processes will be limited by hardware processing power and computing power.

[0067] Option 2: Hierarchical real-time archiving method.

[0068] Please see Figure 1B , Figure 1B This diagram illustrates a hierarchical archiving process for existing technologies. In the diagram, a facial image captured by a camera is first compared with the facial image database of the permanent residents in the district or county where the camera is located. If the comparison is successful, the image is directly archived. If the comparison fails, the image is then compared with the facial image database of the city where the image is located. If the comparison is successful, the image is directly archived. If the comparison fails, the image is then compared with the facial image database of the province where the image is located.

[0069] Scheme 2 can hierarchically archive face images according to the priority order of spatial range from small to large, but it has the following drawbacks:

[0070] Disadvantage 1: The resident population database is updated infrequently, which often results in unsuccessful comparisons. Ultimately, a large number of facial images still need to be compared with the full database, which cannot effectively reduce the amount of computation.

[0071] Disadvantage 2: Grouping cameras by administrative regions can lead to a situation where, if the trajectories between adjacent administrative regions are dense and closely related, a large number of people need to have their images stored in the archive database as permanent residents in adjacent administrative regions at the same time. This requires storing a large amount of duplicate information, putting a huge strain on memory.

[0072] In summary, existing image archiving methods are limited by hardware processing and computing power, resulting in long archiving times, low archiving efficiency, and poor user experience. Therefore, the image archiving method provided in this application aims to solve the above-mentioned technical problems.

[0073] The embodiments of this application will now be described with reference to the accompanying drawings.

[0074] Based on the technical problems mentioned above, and to facilitate understanding of the embodiments of the present invention, the system architecture on which the embodiments of the present invention are based is described below. Please refer to [link to documentation]. Figure 1C , Figure 1C This is a schematic diagram of an image archiving system architecture provided by an embodiment of the present invention. The system adopts a distributed system architecture and can be used to improve image archiving efficiency in large-scale video surveillance scenarios. The system architecture may include an image acquisition device 101 and a server 102.

[0075] Image acquisition device 101 can be multiple independent cameras or multiple devices with cameras. Image acquisition device 101 has basic functions such as video recording or still image capture, and the images can be processed and converted into digital signals recognizable by a processor by the photosensitive component circuit and control component within the camera. For example, in the large-scale video surveillance scenario of this embodiment, since image acquisition device 101 includes multiple cameras, a single server cannot process the facial image data acquired by all cameras. Therefore, multiple cameras can be grouped according to certain preset conditions, such as grouping cameras according to pedestrian trajectory information, and then the facial image data acquired by each camera group can be sent to a designated server for data processing.

[0076] Server 102 possesses high-speed CPU computing power, long-term reliable operation, powerful I / O external data throughput, and better scalability. Generally, servers are capable of responding to service requests, providing services, and ensuring service availability, depending on the services they provide. For example, server 102 can receive facial images acquired by image acquisition device 101 and archive these facial images. However, in the large-scale video surveillance scenario of this embodiment, a single server may not meet the requirements for real-time archiving. Therefore, a distributed system architecture can be adopted, with a single sub-server processing facial images acquired by cameras in the corresponding camera set. It should be noted that the number of servers can be determined according to the business scale in this embodiment, and is not limited here. Optionally, a master server collects the archiving information from all other sub-servers and can perform spatiotemporal partitioning (such as grouping cameras) and regional high-frequency information statistics (such as statistical analysis). Figure 1C (High-frequency archives in region 1, etc.)

[0077] Understandable, Figure 1C The image archiving system architecture described above is merely an exemplary implementation in the embodiments of this application. The image archiving system architecture in the embodiments of this application includes, but is not limited to, the above system architecture.

[0078] Another image archiving system architecture according to an embodiment of the present invention is described below. Please refer to [link / reference needed]. Figure 1D , Figure 1D This is a schematic diagram of another image archiving system architecture provided by an embodiment of the present invention. The image archiving system architecture includes an image acquisition module 201, a hierarchical archiving module 202, a real-time relationship processing module 203, a spatiotemporal division module 204, an archive merging / archive aging module 205, and an information statistics module 206.

[0079] The image acquisition module 201 can be the image acquisition device 101 mentioned above, which can be multiple independent cameras or multiple devices equipped with cameras. The image acquisition module 201 can be used to acquire images such as faces and human bodies, and can also communicate with a server. For example, the image acquisition module 201 can be multiple cameras, which can acquire facial information and then send the facial images to a designated server for real-time archiving.

[0080] The hierarchical archiving module 202 can be stored as program code in the memory of each sub-server. The hierarchical archiving module 202 can receive images sent by the image acquisition module 201, further archive the images, and record image information (such as the image's archive identifier, the camera identifier of the image source, etc.). For example, when the hierarchical archiving module 202 is stored in the memory of sub-server 1, after receiving an image sent by the image acquisition module, it archives the image and records information such as the archive identifier and the camera identifier that captured the image. Optionally, the hierarchical archiving module 202 can send the image information from the corresponding sub-server to the main server, facilitating subsequent updates of camera groups and their corresponding high-frequency archive databases by the main server.

[0081] The real-time relationship processing module 203 can be stored in the main server's memory in the form of program code. The real-time relationship processing module 203 can be used to receive image information (such as image file identifiers, camera identifiers of image sources, etc.) sent by various sub-servers (such as the hierarchical archiving module 202). Furthermore, it can statistically analyze the data processing volume of each camera over a period of time based on the image information, and can also obtain multiple high-frequency pedestrian trajectories based on the image information.

[0082] The spatiotemporal partitioning module 204 can be stored in the main server's memory as program code. In the large-scale video surveillance scenario of this embodiment, since the image acquisition module 201 acquires significantly less data at night, the system's data processing volume decreases. Therefore, the spatiotemporal partitioning module 204 can be invoked at night. The spatiotemporal partitioning module 204 can group multiple cameras based on the data processing volume of each camera or multiple pedestrian trajectories. Simultaneously, the spatiotemporal partitioning module 204 can also send the camera grouping information to each sub-server, and then adjust the correspondence between each sub-server and the image acquisition module 201.

[0083] The file merging / aging module 205 is stored in the main server's memory in the form of program code. The file merging / aging module 205 can be used to merge files that are independently archived by each sub-server, and can also be used to delete files that have not been updated for a long time.

[0084] The information statistics module 206 is stored in the main server's memory in the form of program code. When the information statistics module 206 receives the camera grouping information sent by the spatiotemporal division module 204, it can determine the high-frequency files corresponding to each camera set based on the camera grouping information and image information. Then, it can send these high-frequency files to the corresponding sub-servers, thereby enabling a high-frequency file base library to be obtained for each camera set.

[0085] Understandable, Figure 1D The image archiving system architecture described above is merely an exemplary implementation in the embodiments of this application. The image archiving system architecture in the embodiments of this application includes, but is not limited to, the above system architecture.

[0086] The specific method architecture on which the embodiments of the present invention are based is described below. See also Figure 2A , Figure 2A This is a flowchart illustrating an image archiving method according to an embodiment of this application. The following will describe it in conjunction with the attached diagram. Figure 2A Based on the aforementioned image archiving system architecture, the image archiving method in the embodiments of this application is described. It should be noted that, for a more detailed description of the image archiving method in the embodiments of this application, the corresponding execution entity is described as a server in each process step; however, this does not mean that the embodiments of this application can only perform the corresponding method flow using the described execution entity. In the embodiments of this invention, the image archiving method may include two stages: a camera grouping stage and a real-time image archiving stage.

[0087] The following is a detailed process in the camera grouping stage of this invention:

[0088] Step S201: The server obtains image information based on M cameras.

[0089] Specifically, M cameras can be understood as cameras involved in a video surveillance scenario; image information can be understood as information about images captured by the aforementioned M cameras and which have been archived. For example, image information may include archive file identifiers of multiple archived images, camera identifier information of multiple archived images, and archive time information of multiple archived images, etc.

[0090] Step S202: The server determines N travel trajectories based on the image information.

[0091] The N travel trajectories are associated with L sets of cameras; wherein each of the L sets of cameras includes one or more cameras covered by the N travel trajectories in the M sets of cameras; M is a positive integer; N is a positive integer; and L is a positive integer. Specifically, the N travel trajectories can be understood as multiple pedestrian trajectory information.

[0092] In one possible implementation, the image information includes archive file identifiers for multiple archived images, camera identifier information for the multiple archived images, and archive time information for the multiple archived images; determining N travel trajectories based on the image information includes: determining the N travel trajectories based on the camera identifier information and the archive time information corresponding to each archive file identifier. Specifically, archived images can be understood as archived images captured by the aforementioned M cameras; archive file identifiers can be understood as the file identifier to which the image is archived; camera identifier information can be understood as the identifier of the camera that captured the image; and archive time information can be understood as the time information when the image was archived. For example, as... Figure 2B As shown, Figure 2B This is a flowchart illustrating a process for determining a travel trajectory according to an embodiment of the present invention. The image information in the diagram includes file information for multiple individuals such as Zhang San, Li Si, and Wang Wu. Each file includes camera identification information and archiving time information for the archived image. If, within a certain timeframe, based on this multiple file information, it can be determined that many individuals appeared not only under camera A but also under camera B, and a very small number of individuals appeared under both camera A and camera C, then the trajectory between A and C can be considered a relatively low-frequency travel trajectory, and the trajectory between A and B can be considered a relatively high-frequency travel trajectory. Preferably, the aforementioned trajectory formation can be a high-frequency travel trajectory, thereby... Figure 2B The image contains AB, BC, and CD travel trajectories. It should be noted that in large-scale real-time image archiving scenarios, the image information may include archiving information from tens of millions or hundreds of millions of archived images, thus enabling the determination of multiple travel trajectories. The above example only illustrates the detailed process of determining one travel trajectory; by analogy, other travel trajectories can be determined. In this embodiment of the invention, when the image information includes archive file identifiers, camera identifier information, and archive time information for multiple archived images, images with the same archive file identifier are processed based on the camera identifier information and archive time information to obtain multiple pedestrian trajectories. Then, adjacent cameras are grouped together based on these pedestrian trajectories. By using the method provided in this embodiment of the invention to determine pedestrian trajectories based on relevant information from archived images, pedestrian trajectories that are closer to reality can be obtained, thereby making camera grouping more reasonable, improving image archiving efficiency, and enhancing user experience.

[0093] In one possible implementation, the method further includes: determining the L sets of cameras based on the N travel trajectories. Specifically, the L sets of cameras can be understood as dividing the space into multiple partitions based on the aforementioned multiple travel trajectories, grouping cameras covered in the same partition into a group, thereby obtaining multiple sets of cameras. For example, as... Figure 2C As shown, Figure 2CThis is a schematic diagram of camera grouping provided by an embodiment of the present invention. After confirming multiple travel trajectories, the space can be divided into regions A, B, and C according to the frequency of these trajectories. Cameras A1, B1, C1, and D1 in region A are grouped into one camera set; cameras A2, B2, C2, and D2 in region B are grouped into another camera set; and cameras A3, B3, C3, and D3 in region C are grouped into yet another camera set. It should be noted that in large-scale real-time image archiving scenarios, each camera set may include hundreds or thousands of cameras, which is not limited here. In this application, after confirming multiple pedestrian trajectories, the space can be divided into multiple partitions based on these trajectories. Then, cameras covered by the same partition are grouped together to obtain multiple camera sets. The method provided by this embodiment of the present invention, by grouping multiple cameras, decomposes large-scale real-time image archiving into multiple small-scale real-time image archiving that can be performed simultaneously and in parallel, improving computational efficiency, thereby improving image archiving efficiency and enhancing user experience.

[0094] In one possible implementation, the image information includes camera identification information of multiple archived images; determining the L camera sets based on the N travel trajectories includes: using the camera identification information of the multiple archived images to calculate the data processing volume of each of the M cameras within a first preset time period; and determining the L camera sets based on the data processing volume of each camera and the N travel trajectories. Specifically, the L camera sets can be understood as not only dividing the space into multiple partitions based on the aforementioned travel trajectories, but also grouping cameras covered by densely populated areas into a group based on the data processing volume of each camera within a preset time period, with the boundaries between groups located in sparsely populated areas, thus obtaining multiple camera sets. In this embodiment of the invention, since the image information includes camera identification information of multiple archived images, it is possible to calculate the data processing volume of each camera within a preset time period (e.g., data processing volume within seven days) based on this camera identification information, and then group the multiple cameras based on the data processing volume of each camera within the preset time period and the obtained multiple pedestrian trajectories to obtain multiple camera sets. Because the camera grouping process takes into account both pedestrian trajectories and the amount of data processed by each camera within a preset time period, it is possible to group cameras covering densely populated areas together, with the boundaries between groups located in sparsely populated areas. This avoids excessive duplication of high-frequency archives from different areas, improves image archiving efficiency, and enhances the user experience.

[0095] Step S203: The server generates L high-frequency archive databases based on the L sets of cameras.

[0096] In this system, one high-frequency archive database is matched with one set of cameras; each high-frequency archive database includes multiple high-frequency archives; each high-frequency archive corresponds to an image set of a high-frequency person, and each high-frequency person is a person whose frequency of appearance in the camera set matched by each high-frequency archive database exceeds a preset value or whose appearance frequency is among the highest; the L high-frequency archive databases are used to archive images captured by target cameras, and the target camera is any one of the M cameras. Specifically, the L camera sets can be understood as multiple camera sets obtained by dividing the space based on the above multiple travel trajectories, obtaining multiple partitions, and grouping the cameras covered in the same partition into a group; the L camera sets can also be understood as multiple camera sets obtained by grouping multiple cameras by combining multiple travel trajectories and information such as the data processing volume of each camera; each high-frequency archive database in the L high-frequency archive databases can be understood as a population file that frequently appears in the area covered by the corresponding camera set. For example, Figure 2C As shown in the figure, the cameras are divided into three groups. The camera set corresponding to area A includes camera A1, camera B1, camera C1, and camera D1. Furthermore, based on camera A1, camera B1, camera C1, and camera D1, high-frequency files that frequently appear in this area can be obtained. For example, if people 1 and 2 frequently appear under camera A1, camera B1, camera C1, and camera D1, then the high-frequency file database corresponding to this camera set can include files of people 1 and 2. Similarly, the high-frequency file databases corresponding to the other two camera sets can be obtained.

[0097] Optionally, in large-scale real-time image archiving scenarios, since only a few people travel at night, real-time image archiving is less common at night. Therefore, steps S201 to S203 mentioned above can be performed at night to improve image archiving efficiency and enhance user experience.

[0098] It should be noted that during the initialization phase of the image archiving system, if the aforementioned image information cannot be obtained, the cameras can be grouped according to administrative regions to obtain multiple camera sets. Furthermore, for each camera set, the resident population archive of the area it covers is used as the high-frequency archive base in the above process. Once the aforementioned image information is obtained, the camera groups and their corresponding high-frequency archive bases are updated.

[0099] For example, such as Figure 2D As shown, Figure 2DThis is a schematic diagram of a camera grouping method provided in an embodiment of the present invention. In the diagram, if the image information mentioned above cannot be obtained during the initialization phase, the cameras can be grouped according to the administrative region division method. For example, the space can be divided into region A, region B, etc., according to the administrative region. Region A includes cameras 1 to 4, and region B includes cameras 5 to 8. The camera set of cameras 1 to 4 corresponds to the permanent resident population archive of region A, and the camera set of cameras 5 to 8 corresponds to the permanent resident population archive of region B. The permanent resident population archive of region A includes archives 10 to 13, and region B includes archives 14 to 17. After acquiring the aforementioned image information, multiple travel trajectories can be obtained based on the image information. Furthermore, the cameras are regrouped based on these travel trajectories. This results in cameras 1, 4, 5, and 9 in region A' being grouped together, and cameras 2, 6, 7, and 10 in region B' being grouped together. Simultaneously, statistics show that the high-frequency archive database in region A' includes archives 1, 2, 3, and 4, and the high-frequency archive database in region B' includes archives 1, 3, 5, and 6. The method provided by this invention provides a camera combination and its corresponding high-frequency archive database that are more conducive to real-time image archiving.

[0100] See Figure 3A , Figure 3A This is a flowchart illustrating the real-time image archiving stage of an image archiving method according to an embodiment of this application. The following will combine... Figure 3A The specific process of the image archiving stage in this embodiment of the invention is as follows:

[0101] Step S301: The server acquires a first image of the target object through the target camera.

[0102] Specifically, the target camera can be understood as any one of the M cameras mentioned above; the target object can be understood as a person passing by the target camera; the first image can be understood as a person's face image or a human body image, etc.

[0103] Step S302: The server determines the target camera set corresponding to the target camera from the L camera sets, and determines the target high-frequency archive base corresponding to the target camera set from the L high-frequency archive bases.

[0104] Specifically, the target camera set can be understood as the set of cameras containing the target camera; the target high-frequency file database can be understood as the high-frequency file database corresponding to the target camera set. For example, such as Figure 2DAs shown, when the target camera is camera 1, after camera 1 captures a face image, the high-frequency archive database corresponding to area A′ of camera 1 will be confirmed, and the high-frequency archive database of area A′ will be used as the target high-frequency archive database.

[0105] Step S303: The server determines whether there is a file similar to the first image in the target high-frequency file database.

[0106] Specifically, once the target high-frequency archive database is determined, it can be determined whether there are archives in the target high-frequency archive database that are similar to the first image. Optionally, representative features of each high-frequency archive are extracted and then compared with the first image to determine whether there are archives in the target high-frequency archive database that are similar to the first image.

[0107] Step S304: If it exists, the server archives the first image into the archive and records the archive file identifier, camera identifier information and archive time information of the first image.

[0108] Specifically, if a file similar to the first image exists in the target high-frequency file database, the image is added to that file in real time, and information such as the archived file identifier, camera identification information, and archived time information can be recorded. For example, such as Figure 2D As shown, when the target camera is camera 1, after camera 1 captures a face image, the high-frequency file database corresponding to area A' of camera 1 is confirmed, and the high-frequency file database of area A' is used as the target high-frequency file database. If file 1 in the target file database is similar to the face image, the face image is archived into file 1, and information related to the image is recorded, such as being captured by camera 1, archived into file 1, and the archiving time being 9:00 AM. In this embodiment of the invention, when the target camera captures a face image, it can first determine whether there is a similar file in the high-frequency file database corresponding to the camera set to which the target camera belongs. If so, the captured face image is archived into that file, and information such as the archived file identifier, the target camera identifier, and the archiving time can be recorded, which facilitates subsequent updates to the camera set and its corresponding high-frequency file database. By comparing and archiving the facial images captured by the target camera with the high-frequency archive database obtained above, the captured facial images can be archived in a high probability during the comparison stage with the high-frequency archive database. This reduces the need for comparison with the full archive database, lowers the computational load, improves image archiving efficiency, and enhances the user experience.

[0109] In one possible implementation, the method further includes: if no file similar to the first image exists in the target high-frequency archive database; then determining whether a file similar to the first image exists in the full archive database; the full archive database includes files of people who have appeared under the M cameras; if so, the first image is archived in the archive database, and the archive database identifier, camera identifier information, and archive time information of the first image are recorded. Specifically, the full archive database can be understood as an archive database that includes high-frequency files from the aforementioned multiple high-frequency archive databases. For example, such as... Figure 2D As shown, when the target camera is camera 1, after camera 1 captures a face image, it will confirm the high-frequency file database corresponding to area A′ of camera 1 and use the high-frequency file database of area A′ as the target high-frequency file database. If there is no file similar to the face image in the target file database, it will determine whether there is a similar file in the full file database. If file 5 in the full file database is similar to the face image, the face image will be archived into file 5, and information related to the image will be recorded, such as being captured by camera 1, archived into file 5, and the archiving time being 9:00 AM. In this embodiment of the invention, when there are no similar files in the high-frequency archive database corresponding to the target camera set, the captured facial image can be compared with the full archive database (such as the permanent resident population database of the entire city). If there are similar files in the full archive database, the captured facial image is assigned to that file. At the same time, information such as the archived file identifier, the target camera identifier, and the archived time can be recorded, which facilitates subsequent updates to the camera set and its corresponding high-frequency archive database, improves image archiving efficiency, and enhances user experience.

[0110] In one possible implementation, the method further includes: if no file similar to the first image exists in the full-scale archive database, a new file is created, the first image is archived into the new file, and the archive file identifier, camera identifier information, and archive time information of the first image are recorded. Specifically, when no similar file exists in the full-scale archive database, a new file can be created, and the image can be archived into the newly created file. Simultaneously, the archive file identifier, target camera identifier, and archive time information can be recorded, facilitating subsequent updates to the camera set and its corresponding high-frequency archive database, improving image archiving efficiency, and enhancing user experience. For example, as... Figure 2DAs shown, when the target camera is camera 1, after camera 1 captures a face image, the high-frequency file database corresponding to area A′ of camera 1 is confirmed, and the high-frequency file database of area A′ is used as the target high-frequency file database. If there is no file similar to the face image in the target file database, it is then determined whether there is a similar file in the full file database. If there is also no file similar to the face image in the full file database, a new file can be created, such as file 100. The face image is then archived into file 100, and information related to the image is recorded, such as being captured by camera 1, archived into file 100, and the archiving time being 9:00 AM.

[0111] In one possible implementation, the method further includes: obtaining a target full-scale archive database based on the archived files; determining whether similar files exist in the target full-scale archive database; and merging the similar files if they exist. In this embodiment of the invention, after images captured by a camera are archived, a target full-scale archive database can be obtained based on the archived files, and it can be determined whether similar files exist in the target full-scale archive database. If they exist, the similar files can be merged, thereby reducing the problem of duplicate files, improving image archiving efficiency, and enhancing the user experience.

[0112] In one possible implementation, the method further includes deleting files in the target full-file database that have not been updated within a second preset time period. In this embodiment of the invention, if there are files in the target full-file database that have not been updated for a long time, these files can be deleted from the full-file database. This not only saves storage resources but also reasonably reduces the number of files contained in the full-file database, thereby reducing unnecessary computational overhead during image archiving, improving image archiving efficiency, and enhancing the user experience.

[0113] Optionally, in large-scale real-time image archiving scenarios, since most people only travel during the day, the aforementioned steps S301 to S304 can be performed during the day, thereby improving image archiving efficiency and enhancing user experience.

[0114] In this embodiment of the invention, pedestrian trajectory information can be determined based on image information. Further, based on this pedestrian trajectory information, the space is divided into multiple partitions, and cameras covered by the same partition are grouped together, thus grouping adjacent cameras into the same group. Then, based on the camera grouping information and image information, high-frequency files appearing in each partition are statistically analyzed, resulting in a high-frequency file database. Furthermore, hierarchical image archiving can be performed based on this high-frequency file database. In summary, unlike existing technologies that only group cameras by region and directly use the regional resident population database as the high-frequency file database, this embodiment of the invention divides the space based on pedestrian trajectory information, groups cameras covered by the same partition into groups, and statistically analyzes regional high-frequency files for different partitions. This not only allows for regular maintenance of camera groups in each partition but also regular updates to the high-frequency file database for each partition. This ensures that the high-frequency file database includes population files that frequently appear in the region recently. Therefore, a large number of facial images can be directly archived at the first level of hierarchical image archiving (such as the comparison stage with the high-frequency file database), effectively reducing computational load, improving image archiving efficiency, and enhancing user experience. At the same time, pre-archiving using a regional high-frequency archive database can also improve the probability of images being correctly assigned to the archive and the accuracy of calculations.

[0115] See Figure 3B , Figure 3B This is a flowchart illustrating another image archiving method in an embodiment of this application. The following will combine... Figure 3B The specific process of this invention embodiment is explained as follows:

[0116] Step S401: The main server obtains image information based on M cameras.

[0117] Step S402: The main server determines N travel trajectories based on the image information.

[0118] The N travel trajectories are associated with L camera sets; wherein each of the L camera sets includes one or more cameras covered by the N travel trajectories in the M camera sets; M is an integer greater than 0; N is an integer greater than 0; and L is an integer greater than 0.

[0119] Step S403: The main server generates L high-frequency archive databases based on the L sets of cameras.

[0120] Wherein, one high-frequency file base is matched with one set of cameras; each high-frequency file base includes multiple high-frequency files; each high-frequency file corresponds to a set of images of a high-frequency person, and each high-frequency person is a person whose frequency of appearance in the set of cameras matched with each high-frequency file base exceeds a preset value or whose appearance frequency is among the highest.

[0121] Step S404: The sub-server acquires the first image of the target object through the target camera.

[0122] The target camera is any one of the M cameras.

[0123] Step S405: The sub-server determines the target camera set corresponding to the target camera from the L camera sets, and determines the target high-frequency archive base corresponding to the target camera set from the L high-frequency archive bases.

[0124] Step S406: The sub-server determines whether there is a file similar to the first image in the target high-frequency file database.

[0125] Step S407: If it exists, the sub-server archives the first image into the archive and records the archive file identifier, camera identifier information and archive time information of the first image.

[0126] It should be noted that, for a detailed explanation of steps S401 to S403 in the embodiments of the present invention, please refer to the above-described... Figure 2A Steps S201 to S203 in the above-described embodiments; and steps S404 to S407 in the embodiments of the present invention, can be described in detail in the above-described embodiments. Figure 3A Steps S301 to S304 in the process.

[0127] In this embodiment of the invention, the main server determines pedestrian trajectory information based on image information, further divides the space into multiple partitions based on this pedestrian trajectory information, and groups cameras covered by the same partition together, thereby assigning adjacent cameras to the same group. Then, based on the camera grouping information and image information, high-frequency files appearing in each partition can be counted to obtain a high-frequency file database. Sub-servers can then perform hierarchical image archiving based on this high-frequency file database. In summary, unlike existing technologies that only group cameras by region and directly use the regional resident population database as the high-frequency file database, this embodiment of the invention divides the space based on pedestrian trajectory information, groups cameras covered by the same partition together, and counts regional high-frequency files for different partitions. This not only allows for regular maintenance of camera groups in each partition but also regular updates to the high-frequency file database for each partition, ensuring that the high-frequency file database includes population files that frequently appear in the region. This allows sub-servers to directly archive a large number of face images at the first level of hierarchical image archiving (such as the comparison stage with the high-frequency file database), effectively reducing computational load, improving image archiving efficiency, and enhancing user experience. Meanwhile, the main server is responsible for grouping cameras and compiling a high-frequency archive database, while the sub-servers are responsible for archiving images in real time. This allows multiple sub-servers to archive images simultaneously, achieving distributed computing and improving image archiving efficiency.

[0128] To describe the image archiving method in this application embodiment in more detail, based on the above-described image archiving system architecture and image archiving method, the following will combine... Figure 4A An exemplary description is provided. Figure 4A This is an exemplary flowchart of an image archiving method provided by an embodiment of the present invention. It should be noted that this embodiment of the present invention can be used in a large-scale distributed image archiving system, which can be divided into a daytime processing section and a nighttime processing section. The image archiving system can first use initial real-time archiving to accumulate a certain amount of archived data, and then perform regular real-time archiving. The detailed process is as follows:

[0129] (I) Initialize real-time archiving:

[0130] 1. Since the system initialization process lacks archived information (such as the image information mentioned above), the cameras can be grouped according to administrative divisions. Within each group, the regional resident population archive database should be used to replace the regional high-frequency archive database in the process.

[0131] 2. A facial image obtained in a certain area is compared with the resident population database of the corresponding area. If a match is found in a resident population database, the facial image is added to that database in real time. If no match is found, the image is compared with the entire database. If a match is found, the facial image is added to that database in real time. If no match is found, a new database is created.

[0132] 3. After a face image is archived, the archive ID of the image, the camera ID from which the image originated, and the time the image was acquired will be recorded. Based on this information, the image information will be recorded, organized, and stored for easy statistical analysis later.

[0133] (ii) Regular real-time archiving.

[0134] 1. Based on the stored image information from the past few days, statistical processing is performed, resulting in the following nighttime workflow:

[0135] (1) Calculate the recent data processing volume of each camera, use pedestrian trajectory information to perform data-driven spatiotemporal grouping, refresh the camera grouping, and balance the load of each server; based on which camera image files each pedestrian appears in, the trajectory information of the pedestrian can be identified.

[0136] (2) Collect high-frequency archives in various time and space areas to form a high-frequency archive database for real-time archiving during the day;

[0137] (3) Merge newly created archives in various time and space areas during the day; age archives that have not been updated for a long time.

[0138] 2. Daytime processing procedure:

[0139] (1) The face image obtained in a certain area is compared with the high-frequency archive database of the corresponding area. If a high-frequency archive is matched, the face image is added to the archive in real time.

[0140] (2) If the high-frequency archive database is not matched, the comparison with all archive databases will continue. If a match is found, the face image will be added to the archive in real time. If no match is found, a new archive will be created.

[0141] (3) Use archived information containing file ID, camera ID, time, etc. to store this structured information as image information to facilitate nighttime statistical work.

[0142] Specifically, for Zhang San, who frequently appears in Area A, his facial images are often captured by cameras within that area. After the system has been running for a period of time, Zhang San's image file will appear in the high-frequency file database for that area. When Zhang San's facial image is captured again by a camera in that area, the image will first be compared with the high-frequency file database. If a match is found, the image will be directly added to the corresponding file, and the archiving information will be recorded for subsequent real-time relationship processing and statistics. One day, Zhang San appears in Area B. Zhang San does not frequently appear in this area, and there is no file for him in the high-frequency file database. When his facial image is captured by a camera in this area, it will first be compared with the high-frequency file database. If no match is found, it will be compared with the entire file database. If a match is found, the image will be archived; otherwise, a new file will be created. Li Si was a temporary visitor to this large area. One day, his facial images were captured by cameras in areas A and B. These facial images were first compared with the high-frequency files in their respective areas. If no match was found, they were then compared with the entire archive database. If a match was found, the file was archived directly; otherwise, a new file was created, and the archiving information was recorded. At night, the files in each area were compared with each other, and files that were close in distance were merged.

[0143] In this embodiment of the invention, the main server determines pedestrian trajectory information based on image information, further divides the space into multiple partitions based on this pedestrian trajectory information, and groups cameras covered by the same partition together, thereby assigning adjacent cameras to the same group. Then, based on the camera grouping information and image information, high-frequency files appearing in each partition can be counted to obtain a high-frequency file database. Sub-servers can then perform hierarchical image archiving based on this high-frequency file database. In summary, unlike existing technologies that only group cameras by region and directly use the regional resident population database as the high-frequency file database, this embodiment of the invention divides the space based on pedestrian trajectory information, groups cameras covered by the same partition together, and counts regional high-frequency files for different partitions. This not only allows for regular maintenance of camera groups in each partition but also regular updates to the high-frequency file database for each partition, ensuring that the high-frequency file database includes population files that frequently appear in the region. This allows sub-servers to directly archive a large number of face images at the first level of hierarchical image archiving (such as the comparison stage with the high-frequency file database), effectively reducing computational load, improving image archiving efficiency, and enhancing user experience. Meanwhile, the main server is responsible for grouping cameras and compiling a high-frequency archive database, while the sub-servers are responsible for archiving images in real time. This allows multiple sub-servers to archive images simultaneously, achieving distributed computing and improving image archiving efficiency.

[0144] To describe the image archiving method in the embodiments of this application in more detail, the following is combined with... Figure 4B An exemplary description is provided. Figure 4BThis is an exemplary flowchart of another image archiving method provided by an embodiment of the present invention. It should be noted that this embodiment of the present invention can be used in a small-scale distributed image archiving system. A small-scale real-time archiving system can use a single server for processing, calling different functional modules through multiple processes. These processes correspond to the main server and sub-servers mentioned in the above embodiments, respectively. Since the area is small, the cameras are grouped into a single camera set. The detailed process is as follows:

[0145] From a business process perspective, similar to the above embodiments, the image archiving system can first use initial real-time archiving to accumulate a certain amount of archived data before performing regular real-time archiving. For example, Zhang San frequently appears in the service area and is frequently captured by cameras in the area. After the system has been running for a period of time, Zhang San's image file will appear in the high-frequency file database for that area. When Zhang San is captured by a camera in that area again, the image will first be compared with the high-frequency file database and matched, directly adding it to the corresponding file and recording the archiving information for subsequent real-time relationship processing and statistics. On the other hand, Li Si does not frequently appear in the service area, and there is no file for Li Si in the high-frequency file database. When Li Si occasionally appears in the area and is captured by a camera, the image will first be compared with the high-frequency file database. If it does not match, it will then be compared with the entire file database. If it matches, it will be archived; otherwise, a new file will be created, and the archiving information will be recorded.

[0146] In this embodiment of the invention, pedestrian trajectory information is determined based on image information, and high-frequency archives appearing in the area are statistically analyzed to obtain a high-frequency archive database. Further, hierarchical image archiving can be performed based on this high-frequency archive database. It should be noted that in a small-scale real-time archiving system, the aforementioned L can be 1, meaning that multiple cameras included in the area can be grouped into one camera set. In summary, unlike existing technologies that directly use a regional resident population database as the high-frequency archive database, this embodiment of the invention statistically analyzes regional high-frequency archives based on pedestrian trajectory information. This allows for regular updates to the high-frequency archive databases of each partition, ensuring that the database includes population archives that have recently appeared in the area. This enables the direct archiving of a large number of facial images at the first level of hierarchical image archiving (such as the comparison stage with the high-frequency archive database), effectively reducing computational load, improving image archiving efficiency, and enhancing user experience. Furthermore, pre-archiving using the regional high-frequency archive database also improves the probability of images being correctly assigned to archives and increases computational accuracy.

[0147] The methods of the embodiments of the present invention have been described in detail above, and the related apparatus of the embodiments of the present invention is provided below.

[0148] Please see Figure 5A , Figure 5AThis embodiment of the invention provides a schematic diagram of an image archiving device 50, which may include a first acquisition unit 501, a first processing unit 502, a second processing unit 503, a third processing unit 504, a second acquisition unit 505, a fourth processing unit 506, a fifth processing unit 507, a first archiving unit 508, a second archiving unit 509, a third archiving unit 510, a first merging unit 511, and a first deletion unit 512. The first acquisition unit 501 is used to acquire image information obtained from M cameras.

[0149] The first processing unit 502 is configured to determine N travel trajectories based on the image information; the N travel trajectories are associated with L camera sets; wherein each of the L camera sets includes one or more cameras covered by the N travel trajectories in the M cameras; M is an integer greater than 0; N is an integer greater than 0; and L is an integer greater than 0.

[0150] The second processing unit 503 is used to generate L high-frequency archive bases based on the L sets of cameras; wherein, one high-frequency archive base matches one set of cameras; each high-frequency archive base includes multiple high-frequency archives; each high-frequency archive corresponds to an image set of a high-frequency person, and each high-frequency person is a person whose frequency of appearance in the set of cameras matched in each high-frequency archive base exceeds a preset value or whose appearance frequency is among the highest; the L high-frequency archive bases are used to archive images captured by the target camera, and the target camera is any one of the M cameras.

[0151] In one possible implementation, the image information includes archive file identifiers of multiple archived images, camera identifier information of the multiple archived images, and archive time information of the multiple archived images; the first processing unit 502 is specifically used to: determine the N travel trajectories based on the camera identifier information and the archive time information corresponding to each of the archive file identifiers.

[0152] In one possible implementation, the device further includes a third processing unit 504, configured to determine the L sets of cameras based on the N travel trajectories.

[0153] In one possible implementation, the image information includes camera identification information of multiple archived images; the third processing unit 504 is specifically used to: use the camera identification information of the multiple archived images to count the data processing volume of each of the M cameras within a first preset time period; and determine the L camera set based on the data processing volume of each camera and the N travel trajectories.

[0154] In one possible implementation, the apparatus further includes: a second acquisition unit 505, configured to acquire a first image of a target object using the target camera; a fourth processing unit 506, configured to determine the target camera set corresponding to the target camera from the L camera sets, and to determine the target high-frequency archive database corresponding to the target camera set from the L high-frequency archive databases; a fifth processing unit 507, configured to determine whether there is an archive similar to the first image in the target high-frequency archive database; and a first archiving unit 508, configured to archive the first image into the archive if it exists, and record the archived archive identifier, camera identifier information, and archiving time information of the first image.

[0155] In one possible implementation, the device further includes: a second archiving unit 509, configured to: if there is no file similar to the first image in the target high-frequency archive database; determine whether there is a file similar to the first image in the full archive database; the full archive database includes files of people who have appeared under the M cameras; if there is, archive the first image into the file and record the archived file identifier, camera identifier information and archived time information of the first image.

[0156] In one possible implementation, the device further includes a third archiving unit 510, configured to create a new archive if there is no archive similar to the first image in the full archive database, archive the first image into the new archive, and record the archived archive identifier, camera identifier information and archived time information of the first image.

[0157] In one possible implementation, the apparatus further includes: a first merging unit 511, configured to obtain a target full archive base based on the archived files, determine whether similar files exist in the target full archive base, and merge the similar files if they exist.

[0158] In one possible implementation, the device further includes a first deletion unit 512, used to delete files in the target full-volume archive that have not been updated within a second preset time period.

[0159] It should be noted that the functions of each functional unit in the image archiving device 50 described in this embodiment of the invention are as described above. Figure 2A The relevant descriptions of steps S201-S203 performed in the embodiments of the method described in the text and Figure 3A The relevant descriptions of steps S301-S304 performed in the method embodiment are not repeated here.

[0160] Please see Figure 5B , Figure 5B This embodiment of the invention provides a schematic diagram of an image archiving device 60, which may include a first acquisition unit 601, a first processing unit 602, a second processing unit 603, a second acquisition unit 604, a third processing unit 605, and a first archiving unit 606.

[0161] The first acquisition unit 601 of the main server is used to acquire image information based on M cameras;

[0162] The first processing unit 602 of the main server is used to determine N travel trajectories based on the image information; the N travel trajectories are associated with L camera sets; wherein, each of the L camera sets includes one or more cameras covered by the N travel trajectories in the M cameras; M is an integer greater than 0; N is an integer greater than 0; and L is an integer greater than 0.

[0163] The second processing unit 603 of the main server is used to generate L high-frequency file bases based on the L camera sets; wherein, one high-frequency file base matches one camera set; each high-frequency file base includes multiple high-frequency files; each high-frequency file corresponds to an image set of a high-frequency person, and each high-frequency person is a person whose frequency of appearance in the camera set matched by each high-frequency file base exceeds a preset value or whose appearance frequency is among the highest.

[0164] The second acquisition unit 604 of the sub-server is used to acquire a first image of the target object through the target camera; the target camera is any one of the M cameras.

[0165] The third processing unit 605 of the sub-server is used to determine the target camera set corresponding to the target camera from the L camera sets, and to determine the target high-frequency archive base corresponding to the target camera set from the L high-frequency archive bases.

[0166] The first archiving unit 606 of the sub-server is used to determine whether there is a file similar to the first image in the target high-frequency file base; if there is, the sub-server archives the first image into the file and records the archive file identifier, camera identifier information and archive time information of the first image.

[0167] It should be noted that the functions of each functional unit in the image archiving device 60 described in this embodiment of the invention are as described above. Figure 3B The relevant descriptions of steps S401-S407 performed in the method embodiment are not repeated here.

[0168] This invention provides a chip system, characterized in that the chip system includes at least one processor, a memory, and an interface circuit, wherein the memory, the interface circuit, and the at least one processor are interconnected via circuits, and the at least one memory stores instructions; when the instructions are executed by the processor, the above... Figure 2A and Figure 3A Any one of the steps performed in an embodiment of an image archiving method is implemented.

[0169] This invention provides a chip system, characterized in that the chip system includes at least one processor, a memory, and an interface circuit, wherein the memory, the interface circuit, and the at least one processor are interconnected via circuits, and the at least one memory stores instructions; when the instructions are executed by the processor, the above... Figure 3B Any one of the steps performed in an embodiment of an image archiving method is implemented.

[0170] This invention provides a computer storage medium, characterized in that the computer storage medium stores a computer program, which, when executed by a processor, implements the above-described... Figure 2A and Figure 3A Any step performed in an embodiment of an image archiving method.

[0171] This invention provides a computer storage medium, characterized in that the computer storage medium stores a computer program, which, when executed by a processor, implements the above-described... Figure 3B Any step performed in an embodiment of an image archiving method.

[0172] This invention provides a computer program, characterized in that the computer program includes instructions, which, when executed by a computer, cause the computer to perform the aforementioned... Figure 2A and Figure 3A Any step performed in an embodiment of an image archiving method.

[0173] This invention provides a computer program, characterized in that the computer program includes instructions, which, when executed by a computer, cause the computer to perform the aforementioned... Figure 3B Any step performed in an embodiment of an image archiving method.

[0174] This application provides a terminal device that has the functions described above. Figure 2A and Figure 3AThis refers to the functionality of any of the image archiving methods provided. This functionality can be implemented in hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described functionality.

[0175] This application provides a terminal device that has the functions described above. Figure 3B This refers to the functionality of any of the image archiving methods provided. This functionality can be implemented in hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described functionality.

[0176] This application provides a smart device that has the functions described above. Figure 2A and Figure 3A This refers to the functionality of any of the image archiving methods provided. This functionality can be implemented in hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described functionality.

[0177] This application provides a smart device that has the functions described above. Figure 3B This refers to the functionality of any of the image archiving methods provided. This functionality can be implemented in hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described functionality.

[0178] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0179] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0180] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0181] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0182] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0183] If the integrated units described above are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which can be a personal computer, server, or network device, specifically a processor in the computer device) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium may include various media capable of storing program code, such as a USB flash drive, portable hard drive, magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM).

[0184] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for image archiving, characterized in that, The method includes: Acquire image information from M cameras; Based on the image information, N travel trajectories are determined; the N travel trajectories are associated with L camera sets; wherein, each of the L camera sets includes one or more cameras covered by the N travel trajectories in the M camera sets; M is an integer greater than 0; N is an integer greater than 0; L is an integer greater than 0. L high-frequency archive databases are generated based on the L sets of cameras; wherein, one high-frequency archive database is matched with one set of cameras; each high-frequency archive database includes multiple high-frequency archives; each high-frequency archive corresponds to a set of images of a high-frequency person, and each high-frequency person is a person whose frequency of appearance in the set of cameras matched with each high-frequency archive database exceeds a preset value or whose appearance frequency is among the highest; the L high-frequency archive databases are used to archive images captured by the target camera, and the target camera is any one of the M cameras.

2. The method as described in claim 1, characterized in that, The image information includes archive file identifiers for multiple archived images, camera identifier information for the multiple archived images, and archive time information for the multiple archived images; The step of determining N travel trajectories based on the image information includes: Based on the camera identification information and the archiving time information corresponding to each archived file identifier, the N travel trajectories are determined.

3. The method as described in claim 1, characterized in that, The method further includes: Based on the N travel trajectories, the L sets of cameras are determined.

4. The method as described in claim 3, characterized in that, The image information includes camera identification information for multiple archived images; determining the L camera sets based on the N travel trajectories includes: Using the camera identification information of the multiple archived images, the data processing volume of each of the M cameras is calculated within a first preset time period; Based on the data processing volume of each camera and the N travel trajectories, the set of L cameras is determined.

5. The method as described in claim 1, characterized in that, The method further includes: The target camera captures a first image of the target object; The target camera set corresponding to the target camera is determined from the L sets of cameras, and the target high-frequency archive base corresponding to the target camera set is determined from the L high-frequency archive bases; Determine whether there is a file similar to the first image in the target high-frequency file database; If it exists, the first image is archived in the archive, and the archive file identifier, camera identifier information and archive time information of the first image are recorded.

6. The method as described in claim 5, characterized in that, The method further includes: If there is no file similar to the first image in the target high-frequency file database; Then determine whether there is a file similar to the first image in the full archive database; the full archive database includes files of people who have appeared under the M cameras; If it exists, the first image is archived in the archive, and the archive file identifier, camera identifier information and archive time information of the first image are recorded.

7. The method as described in claim 6, characterized in that, The method further includes: If there are no files similar to the first image in the full archive database; Then a new file is created, the first image is archived in the new file, and the archive file identifier, camera identifier information and archive time information of the first image are recorded.

8. The method according to any one of claims 5-7, characterized in that, The method further includes: Based on the archived files, a target full-scale archive base is obtained, and it is determined whether similar files exist in the target full-scale archive base. If they exist, similar files will be merged.

9. The method as described in claim 8, characterized in that, The method further includes: Delete the files in the target full archive that have not been updated within a second preset time period.

10. A method for image archiving, characterized in that, The method includes: The main server acquires image information from M cameras; The main server determines N travel trajectories based on the image information; the N travel trajectories are associated with L camera sets; wherein each of the L camera sets includes one or more cameras covered by the N travel trajectories in the M camera sets; M is an integer greater than 0; N is an integer greater than 0; L is an integer greater than 0. The main server generates L high-frequency file bases based on the L camera sets; wherein, one high-frequency file base matches one camera set; each high-frequency file base includes multiple high-frequency files; each high-frequency file corresponds to an image set of a high-frequency person, and each high-frequency person is a person whose frequency of appearance in the camera set matched in each high-frequency file base exceeds a preset value or whose appearance frequency is among the highest. The sub-server acquires a first image of the target object through the target camera; the target camera is any one of the M cameras. The sub-server determines the target camera set corresponding to the target camera from the L camera sets, and determines the target high-frequency archive base corresponding to the target camera set from the L high-frequency archive bases; The sub-server determines whether there is a file similar to the first image in the target high-frequency file database; If it exists, the sub-server archives the first image into the archive and records the archive file identifier, camera identifier information and archive time information of the first image.

11. An image archiving apparatus, characterized in that, The device includes: The first acquisition unit is used to acquire image information based on M cameras; The first processing unit is configured to determine N travel trajectories based on the image information; the N travel trajectories are associated with L camera sets; wherein each of the L camera sets includes one or more cameras covered by the N travel trajectories in the M camera sets; M is an integer greater than 0; N is an integer greater than 0; and L is an integer greater than 0. The second processing unit is used to generate L high-frequency archive databases based on the L camera sets; wherein, one high-frequency archive database matches one camera set; each high-frequency archive database includes multiple high-frequency files; each high-frequency file corresponds to an image set of a high-frequency person, and each high-frequency person is a person whose frequency of appearance in the camera set matched by each high-frequency archive database exceeds a preset value or whose appearance frequency is among the highest; the L high-frequency archive databases are used to archive images captured by the target camera, and the target camera is any one of the M cameras.

12. The apparatus as claimed in claim 11, characterized in that, The image information includes archive file identifiers for multiple archived images, camera identifier information for the multiple archived images, and archive time information for the multiple archived images; the first processing unit is specifically used for: Based on the camera identification information and the archiving time information corresponding to each archived file identifier, the N travel trajectories are determined.

13. The apparatus as claimed in claim 11, characterized in that, The device further includes: The third processing unit is used to determine the set of L cameras based on the N travel trajectories.

14. The apparatus as claimed in claim 13, characterized in that, The image information includes camera identification information for multiple archived images; the third processing unit is specifically used for: Using the camera identification information of the multiple archived images, the data processing volume of each of the M cameras is calculated within a first preset time period; Based on the data processing volume of each camera and the N travel trajectories, the set of L cameras is determined.

15. The apparatus as claimed in claim 11, characterized in that, The device further includes: The second acquisition unit is used to acquire a first image of the target object through the target camera; The fourth processing unit is used to determine the target camera set corresponding to the target camera from the L camera sets, and to determine the target high-frequency archive base corresponding to the target camera set from the L high-frequency archive bases; The fifth processing unit is used to determine whether there are files similar to the first image in the target high-frequency archive database; The first archiving unit is used to archive the first image into the archive if it exists, and to record the archive file identifier, camera identifier information and archiving time information of the first image.

16. The apparatus as claimed in claim 15, characterized in that, The device further includes: The second archiving unit is used to determine whether there is a file similar to the first image in the full archive if there is no file similar to the first image in the target high-frequency archive database. The full archive database includes files of people who have appeared under the M cameras. If there is a file similar to the first image, the first image is archived in the archive database, and the archived file identifier, camera identifier information and archived time information of the first image are recorded.

17. The apparatus as claimed in claim 16, characterized in that, The device further includes: The third archiving unit is used to create a new archive if there is no archive similar to the first image in the full archive database, archive the first image into the new archive, and record the archive identifier, camera identifier information and archive time information of the first image.

18. The apparatus according to any one of claims 15-17, characterized in that, The device further includes: The first merging unit is used to obtain a target full archive base based on the archived files, and to determine whether there are similar files in the target full archive base; if so, the similar files are merged.

19. The apparatus as claimed in claim 18, characterized in that, The device further includes: The first deletion unit is used to delete files in the target full-volume archive that have not been updated within a second preset time.

20. An image archiving apparatus, characterized in that, The device includes: The first acquisition unit of the main server is used to acquire image information based on M cameras; The first processing unit of the main server is used to determine N travel trajectories based on the image information; the N travel trajectories are associated with L camera sets; wherein, each of the L camera sets includes one or more cameras covered by the N travel trajectories in the M cameras; M is an integer greater than 0; N is an integer greater than 0; and L is an integer greater than 0. The second processing unit of the main server is used to generate L high-frequency archive bases based on the L camera sets; wherein, one high-frequency archive base matches one camera set; each high-frequency archive base includes multiple high-frequency archives; each high-frequency archive corresponds to an image set of a high-frequency person, and each high-frequency person is a person whose frequency of appearance in the camera set matched by each high-frequency archive base exceeds a preset value or whose appearance frequency is among the highest. The second acquisition unit of the sub-server is used to acquire a first image of the target object through the target camera; the target camera is any one of the M cameras. The third processing unit of the sub-server is used to determine the target camera set corresponding to the target camera from the L camera sets, and to determine the target high-frequency archive base corresponding to the target camera set from the L high-frequency archive bases. The first archiving unit of the sub-server is used to determine whether there is a file similar to the first image in the target high-frequency file base; if there is, the sub-server archives the first image into the file and records the archive file identifier, camera identifier information and archive time information of the first image.

21. A chip system, characterized in that, The chip system includes at least one processor, a memory, and an interface circuit. The memory, the interface circuit, and the at least one processor are interconnected via lines. The at least one memory stores instructions. When the instructions are executed by the processor, the method described in any one of claims 1-9 is implemented.

22. A chip system, characterized in that, The chip system includes at least one processor, a memory, and an interface circuit. The memory, the interface circuit, and the at least one processor are interconnected via lines. The at least one memory stores instructions. When the instructions are executed by the processor, the method of claim 10 is implemented.

23. A computer storage medium, characterized in that, The computer storage medium stores a computer program that, when executed by a processor, implements the method described in any one of claims 1-9.

24. A computer storage medium, characterized in that, The computer storage medium stores a computer program that, when executed by a processor, implements the method described in claim 10.

25. A computer program product, characterized in that, The computer program product includes instructions that, when executed by a computer, cause the computer to perform the method as described in any one of claims 1-9.

26. A computer program product, characterized in that, The computer program product includes instructions that, when executed by a computer, cause the computer to perform the method as described in claim 10.

Images