A method, device and electronic device for clustering target objects

Abstract

The present application relates to a target object clustering method, device and electronic device. The method includes: obtaining the target object picture frame numbers corresponding to multiple target object archives respectively; calculating the first target object acquisition points corresponding to each target object archive in a specified time period respectively according to the target object picture frame numbers; screening out at least one designated object acquisition point from each first target object acquisition point, and determining the designated object archives corresponding to each designated object acquisition point respectively; adjusting the clustering thresholds corresponding to each designated object archive respectively, and clustering the target object pictures corresponding to each designated object archive respectively according to the adjusted clustering thresholds. Through the above method, the designated object archives corresponding to the target objects that frequently appear in the target area are determined, and the clustering thresholds corresponding to the designated object archives are adjusted, which helps to improve the recall rate corresponding to each target object archive in the target object clustering process.

Description

Technical Field

[0001] The present application relates to the technical field of image processing target objects, and in particular to a target object clustering method, device and electronic equipment. Background Art

[0002] With the rapid development of society, clustering target objects has become one of the key means of target locking. Based on the target object images collected by cameras in the area over a period of time, deep learning technology is used to identify the target object images, extract the target object features in the target object images, cluster similar target object images, form a target object file corresponding to each target object, and lock the target object based on the trajectory information of the target object in the target object file.

[0003] At the present stage, most of the above processes are based on deep learning technology to extract target object features in target object pictures, and then cluster target object pictures whose similarity values ​​are greater than the clustering threshold according to the target object features. However, due to the acquisition angle of the target object pictures, picture clarity and other reasons, the similarity of target object pictures corresponding to the same target object may be slightly lower than the clustering threshold, resulting in missed clustering of target object pictures. If the clustering threshold is lowered for target object pictures corresponding to all target objects according to a unified standard, it will cause target object picture clustering errors, resulting in low accuracy of target object clustering. Summary of the invention

[0004] The present application provides a target object clustering method, device and electronic device, which determine a specified object file from each target object file corresponding to a target area, and adjust the clustering threshold corresponding to the specified object file, which helps to improve the recall rate corresponding to each target object file during the target object clustering process, thereby improving the accuracy of target object clustering.

[0005] In a first aspect, the present application provides a target object clustering method, the method comprising:

[0006] Obtaining target object picture frame numbers corresponding to a plurality of target object files respectively, wherein the target object pictures in the plurality of target object files are pictures collected within a target area and within a preset time period;

[0007] Calculate the first target object acquisition score corresponding to each target object file in a specified time period according to the number of frames of each target object picture, wherein the specified time period includes multiple preset time periods;

[0008] Selecting at least one designated object collection score from each first target object collection score, and determining designated object profiles corresponding to each designated object collection score;

[0009] The clustering thresholds corresponding to the designated object files are adjusted, and the target object images corresponding to the designated object files are clustered according to the adjusted clustering thresholds.

[0010] Through the above method, the designated object files where the target object frequently appears are determined in each target object file corresponding to the target area, and the clustering threshold corresponding to the designated object file is adjusted, which helps to improve the recall rate corresponding to each target object file in the target object clustering process, thereby improving the accuracy of target object clustering.

[0011] In a possible design, obtaining the target object picture frame numbers corresponding to the plurality of target object files respectively includes:

[0012] Acquire multiple target object files corresponding to the target object images collected within the target area and within the specified time period;

[0013] Respectively extracting first trajectory information corresponding to the target object in each target object file within each preset time period;

[0014] According to each piece of first trajectory information, the number of target object picture frames corresponding to each target object file is obtained.

[0015] Through the above method, the number of target object picture frames in each target object file in the target area and the specified time period is obtained, and the number of target object picture frames is used to determine the target object files in which the target object frequently appears in the target area.

[0016] In a possible design, respectively extracting the first trajectory information corresponding to the target object in each target object file within each preset time period includes:

[0017] According to the collection points corresponding to the target object pictures in each target object file, second trajectory information corresponding to each target object file is obtained;

[0018] The longitude and latitude information of the collection points corresponding to each second track information is converted according to a preset code to obtain third track information;

[0019] The target object images in each of the third track information are deduplicated to obtain the first track information.

[0020] The first trajectory information obtained by the above method eliminates the influence of repeatedly collected target object pictures, which helps to improve the accuracy of the subsequent determination of the number of target object picture frames in the target object file, thereby improving the recall rate of each target object file in the target object clustering process.

[0021] In a possible design, the first target object acquisition score corresponding to each target object file in a specified time period is calculated according to the number of frames of each target object picture, including:

[0022] According to the number of frames of each target object picture, a plurality of second target object acquisition points corresponding to each target object file are calculated respectively, wherein the same target object file contains pictures acquired in a plurality of preset time periods, and one of the preset time periods corresponds to one second target object acquisition point;

[0023] Setting weight values ​​for each second target object collection score corresponding to the same target object profile, respectively, to obtain third target object collection scores corresponding to a plurality of preset time periods;

[0024] The multiple third target object acquisition scores corresponding to the same target object file are summed up to obtain the first target object acquisition score corresponding to each target object file in the specified time period.

[0025] Through the above method, the target object acquisition score corresponding to each target object file can be calculated, and the target object acquisition score is used to screen out the target object files corresponding to the target objects that frequently appear in the target area.

[0026] In a possible design, the step of selecting at least one designated object collection score from each first target object collection score and determining designated object profiles corresponding to each designated object collection score includes:

[0027] Sort the collection scores of the first target objects from largest to smallest;

[0028] The first M first target object acquisition scores among the sorted plurality of first target object acquisition scores are used as designated object acquisition scores;

[0029] The designated object files corresponding to the designated object collection points are determined from the plurality of target object files.

[0030] Through the above method, the target object files corresponding to the target objects that frequently appear in the target area are determined in each target object file.

[0031] In a possible design, adjusting the clustering thresholds corresponding to the respective designated object profiles includes:

[0032] Determine the attenuation values ​​corresponding to the respective designated object files according to the arrangement order of the first target object acquisition points corresponding to the respective designated object files;

[0033] According to the attenuation value, the clustering threshold corresponding to the designated object file is adjusted.

[0034] By using the above method, the clustering threshold corresponding to the specified object file is adjusted, which helps to improve the recall rate of each target object file in the target area corresponding target object clustering process.

[0035] In a second aspect, the present application provides a target object clustering device, the device comprising:

[0036] An acquisition module, used to acquire target object picture frame numbers corresponding to a plurality of target object files, wherein the target object pictures in the plurality of target object files are pictures collected within a target area and within a preset time period;

[0037] A calculation module, used to calculate the first target object acquisition score corresponding to each target object file in a specified time period according to the number of frames of each target object picture, wherein the specified time period includes a plurality of preset time periods;

[0038] A screening module, used to screen out at least one designated object collection score from each first target object collection score, and determine the designated object files corresponding to each designated object collection score;

[0039] The clustering module is used to adjust the clustering thresholds corresponding to the respective designated object files, and cluster the target object images corresponding to the respective designated object files according to the adjusted clustering thresholds.

[0040] In a possible design, the acquisition module is specifically used for:

[0041] Acquire multiple target object files corresponding to the target object images collected within the target area and within the specified time period;

[0042] Respectively extracting first trajectory information corresponding to the target object in each target object file within each preset time period;

[0043] According to each piece of first trajectory information, the number of target object picture frames corresponding to each target object file is obtained.

[0044] In a possible design, the acquisition module is further used for:

[0045] According to the collection points corresponding to the target object pictures in each target object file, second trajectory information corresponding to each target object file is obtained;

[0046] The longitude and latitude information of the collection points corresponding to each second track information is converted according to a preset code to obtain third track information;

[0047] The target object images in each of the third track information are deduplicated to obtain the first track information.

[0048] In a possible design, the computing module is specifically used for:

[0049] According to the number of frames of each target object picture, a plurality of second target object acquisition points corresponding to each target object file are calculated respectively, wherein the same target object file contains pictures acquired in a plurality of preset time periods, and one of the preset time periods corresponds to one second target object acquisition point;

[0050] Setting weight values ​​for each second target object collection score corresponding to the same target object profile, respectively, to obtain third target object collection scores corresponding to a plurality of preset time periods;

[0051] The multiple third target object acquisition scores corresponding to the same target object file are summed up to obtain the first target object acquisition score corresponding to each target object file in the specified time period.

[0052] In a possible design, the screening module is specifically used for:

[0053] Sort the collection scores of the first target objects from largest to smallest;

[0054] The first M first target object acquisition scores among the sorted plurality of first target object acquisition scores are used as designated object acquisition scores;

[0055] The designated object files corresponding to the designated object collection points are determined from the plurality of target object files.

[0056] In a possible design, the clustering module is specifically used for:

[0057] Determine the attenuation values ​​corresponding to the respective designated object files according to the arrangement order of the first target object acquisition points corresponding to the respective designated object files;

[0058] According to the attenuation value, the clustering threshold corresponding to the designated object file is adjusted.

[0059] In a third aspect, the present application provides an electronic device, including:

[0060] Memory, used to store computer programs;

[0061] The processor is used to implement the above-mentioned target object clustering method steps when executing the computer program stored in the memory.

[0062] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned target object clustering method steps are implemented.

[0063] Based on the above target object clustering method, the designated object files where the target objects frequently appear are determined in each target object file corresponding to the target area, and the clustering threshold corresponding to the designated object file is adjusted, which helps to improve the recall rate corresponding to each target object file in the target object clustering process, thereby improving the accuracy of target object clustering corresponding to the target area.

[0064] The technical effects that can be achieved in each of the above-mentioned second to fifth aspects and each of the above-mentioned aspects refer to the technical effects that can be achieved in the above-mentioned first aspect or various possible schemes in the first aspect, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 A flow chart of a target object clustering method provided in this application;

[0066] Figure 2 A schematic diagram of the structure of a target object clustering device provided in this application;

[0067] Figure 3 A schematic diagram of the structure of an electronic device provided in this application. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The specific operating methods in the method embodiments can also be applied to device embodiments or system embodiments. It should be noted that in the description of the present application, "multiple" is understood as "at least two". "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A is connected to B, which can represent: A is directly connected to B and A is connected to B through C. In addition, in the description of the present application, words such as "first" and "second" are only used to distinguish the purpose of description, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying order.

[0069] The embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0070] Most of the existing target object clustering technologies are based on deep learning technology, which extracts target object features from target object images, and then clusters target object images with similarity values ​​greater than a clustering threshold based on the target object features. However, due to the capture angle, image clarity and other reasons in different situations, the similarity of target object images corresponding to the same target object may sometimes be slightly lower than the clustering threshold, resulting in missed clustering. If the clustering threshold is lowered for target object images corresponding to all target objects according to a unified standard, clustering errors will occur, resulting in low clustering accuracy.

[0071] In order to solve the above problems, the present application provides a target object clustering method, which determines the designated object files where the target object frequently appears in each target object file corresponding to the target area, and adjusts the clustering threshold corresponding to the designated object file, which helps to improve the recall rate corresponding to each target object file in the target object clustering process, thereby improving the accuracy of the target object clustering corresponding to the target area. Among them, the method and device described in the embodiments of the present application are based on the same technical concept. Since the principles of the problems solved by the method and the device are similar, the embodiments of the device and the method can refer to each other, and the repeated parts will not be repeated.

[0072] like Figure 1 FIG. 1 is a flowchart of a target object clustering method provided in the present application, which specifically includes the following steps:

[0073] S11, obtaining target object picture frame numbers corresponding to a plurality of target object files respectively;

[0074] In an embodiment of the present application, before performing target object clustering, multiple target object files corresponding to target object images collected within a specified time period in the target area are first obtained, wherein the target object can be a person or a vehicle, which is not specifically limited here. The target area can be a geohash block, and the specified time period includes multiple preset time periods. If the specified time period is the previous 30 days, the preset time period can be set to yesterday, the day before yesterday, etc., or to the tth day in the previous 30 days, where t is an integer greater than or equal to 1 and less than or equal to 30.

[0075] Furthermore, the first trajectory information corresponding to the target object in each target object file in each preset time period is extracted respectively. The specific extraction method may be:

[0076] According to the collection points corresponding to each target object picture in each target object file, the second trajectory information corresponding to each target object file is obtained, wherein the collection point can be the target object picture collection card port. i The second trajectory information corresponding to the target object j days before the current date is D ij , then Dij The specific formula is as follows:

[0077]

[0078] In formula (1), C i The camera head corresponding to the i-th target object image acquisition, where i = 1,…,n.

[0079] Next, the longitude and latitude information of the collection points corresponding to each second trajectory information is converted according to a preset code to obtain the third trajectory information, wherein the preset code may be a geohash code, such as geohash7, and the geohash code is used to convert the second trajectory information D ij After conversion, the third trajectory information obtained is D′ ij , D′ ij The specific formula is as follows:

[0080]

[0081] In formula (2), G i For bayonet C i The corresponding geohash code.

[0082] Furthermore, the target object images in each third track information are deduplicated to obtain the first track information. ij There is C i =C i+1 When i There is a repeated target object image collection at , so the target object image collected at the i+1th time is deleted, and the archive trajectory D′ ij After deduplication, the new trajectory information D″ is obtained ij The specific formula is as follows:

[0083]

[0084] Through the above method, the first trajectory information corresponding to the target object in each target object file in each preset time period can be obtained. Further, according to each first trajectory information, the number of target object picture frames corresponding to each target object file is obtained, wherein the target object pictures in the target object file are pictures collected within the target area and within the preset time period, for example, the number of target object picture frames collected every day within a geohash block.

[0085] S12, calculating the first target object acquisition score corresponding to each target object file in the specified time period according to the number of frames of each target object picture;

[0086] In the embodiment of the present application, after obtaining the number of target object picture frames collected in the target area and the specified time period in the target object file, multiple second target object collection points corresponding to each target object file are calculated respectively according to the number of target object picture frames, wherein the same target object file contains pictures collected in multiple preset time periods, and one preset time period corresponds to one second target object collection point, and the target object collection point indicates the frequency of the target object in the target object file appearing in the target area. The formula for calculating the second target object collection point is as follows:

[0087]

[0088] In formula (4), a is the number of target object image frames collected in the target area and in the jth preset time period in the target object file, S j Collect points for the second target object corresponding to the jth preset time period. i On day j, in geohash block G k If a target object images are collected a times, then formula (4) can express the archive D i On day j, in geohash block G k The corresponding second target object collects points.

[0089] Furthermore, weight values ​​are set for each second target object collection score corresponding to the same target object profile, and third target object collection scores corresponding to multiple preset time periods are obtained, wherein the setting of the weight value is related to the preset time period corresponding to the second target object collection score. If the preset time period is closer to the current time, the larger the weight value is set, and the second target object collection score S is higher. j The corresponding weight value can be set to (1-j / N), then S j The corresponding third target object acquisition score is (1-j / N)•S j .

[0090] Furthermore, the multiple third target object collection scores corresponding to each target object file are summed to obtain the first target object collection score corresponding to each target object file in the specified time period. The calculation formula corresponding to the first target object collection score is:

[0091]

[0092] In formula (4), S all Represents the second target object acquisition score S j The corresponding first target object collection points, N is the number of preset time periods included in the specified time period, and the value of N is an integer greater than 1.

[0093] Through the above method, the first target object collection points corresponding to each target object file in the specified time period can be obtained, and during the calculation process, the weight value is set according to the relationship between the preset time period and the current time, and the proportion of the second target object collection points corresponding to the preset time period closer to the current time is increased.

[0094] S13, selecting at least one designated object collection score from each first target object collection score, and determining the designated object files corresponding to each designated object collection score;

[0095] In an embodiment of the present application, after obtaining the first target object acquisition score corresponding to each target object archive, the first target object acquisition score is sorted from large to small, and the first M first target object acquisition scores among the sorted multiple first target object acquisition scores are used as designated object acquisition scores, where M is an integer greater than or equal to 1. The M first target object acquisition scores can be the top y%, the top 2y%, or the top 3y%, where the value of y is generally 5, and can also be adjusted according to actual conditions, and is not specifically limited here. Finally, the designated object archives corresponding to each designated object acquisition score are determined in the multiple target object archives, and these designated object archives are the target object archives corresponding to the target objects that frequently appear in the target area.

[0096] S14, adjusting the clustering thresholds corresponding to the designated object files respectively, and clustering the target object images corresponding to the designated object files according to the adjusted clustering thresholds.

[0097] In the embodiment of the present application, after determining the designated object files corresponding to the target area, the attenuation values ​​corresponding to the designated object files are determined according to the arrangement order of the first target object acquisition scores corresponding to the designated object files. For example, if the first target object acquisition score corresponding to the target object is in the first y%, the attenuation value can be determined as 3x, where x is generally 0.01, and of course it can be adjusted according to the specific situation; if the first target object acquisition score corresponding to the target object is in the first y%, then the attenuation value can be determined as 2x; if the first target object acquisition score corresponding to the target object is in the first y%, then the attenuation value can be determined as x, that is, the first target object acquisition score with the first arrangement order has the larger attenuation value. After determining the attenuation values ​​corresponding to the designated object files, the target object clustering thresholds corresponding to the designated object files are adjusted according to the respective attenuation values.

[0098] Further, the target object collection information corresponding to the current time period of the target area is obtained, and each target object file after adjusting the target object clustering threshold is used to cluster the target object collection information.

[0099] For example, first cluster the target object files and target object collection information with the target object clustering threshold reduced by 3x, then cluster the target object files and target object collection information with the target object clustering threshold reduced by 2x, then cluster the target object files and target object collection information with the target object clustering threshold reduced by x, and finally cluster the remaining target object files and target object collection information, and finally obtain the target object clustering results for the day.

[0100] In a possible embodiment, a key-value pair may be generated for the target object file corresponding to the target area after the target object is clustered and the adjusted target object clustering threshold. For example, the target object file D i In the geohash block G k The key-value pair after the target object clustering threshold drops by x is {(G k ,D i ):x}, finally, according to the key-value pair, the frequently appearing specified object files in the target area can be determined, and the clustering threshold corresponding to the specified object files can be directly obtained.

[0101] Based on the target object clustering method, the designated object files where the target object frequently appears are determined in each target object file corresponding to the target area, and the clustering threshold corresponding to the designated object file is adjusted, which helps to improve the recall rate corresponding to each target object file in the target object clustering process.

[0102] Furthermore, in order to elaborate on a target object clustering method provided by the present application in more detail, the method provided by the present application is described in detail below through a specific application scenario. Taking the clustering of vehicles as an example, first obtain the archives of all clustered vehicles at a certain intersection in the previous 30 days, and obtain the trajectory information corresponding to the vehicle pictures of each day in each archive, and then, based on the longitude and latitude information corresponding to the acquisition point of each vehicle picture in the archive, convert the trajectory information corresponding to the vehicle picture through geohash coding to obtain new trajectory information. Next, deduplicate the new trajectory information according to the set deduplication rules, remove the duplicate pictures collected continuously, and obtain the deduplicated trajectory information.

[0103] Furthermore, based on the deduplicated trajectory information, the number of times the vehicle in each archive was collected in each geohash block was calculated for each day in the previous 30 days, and the corresponding collected points for each archive in each geohash block were calculated based on the number of times collected. Next, the collected points for each archive in each geohash block every day were accumulated to obtain the total collected points for each archive in each geohash block in the previous 30 days.

[0104] Furthermore, the total collected points of each archive corresponding to each geohash block in the previous 30 days are sorted in order from large to small, and the top 5%, top 10% and top 15% of the total collected points of each geohash block in the previous 30 days are obtained respectively. These archives are used as frequently appearing archives corresponding to the geohash block. Then, the vehicle clustering threshold of these archives in the corresponding geohash is lowered, and a key-value pair is generated for each archive at the clustering threshold that drops at each geohash.

[0105] Furthermore, the vehicle image information corresponding to all the acquisition checkpoints of the intersection on that day is obtained, and the latitude and longitude information of the acquisition checkpoints in the vehicle image information is converted according to the geohash code, and then the frequently appearing files in each geohash block are found according to the key-value pairs corresponding to the vehicle files of the previous 30 days, and then the frequently appearing files are clustered according to the order of the clustering threshold adjustment amount from large to small, and the vehicle clustering results of the day are obtained. Finally, the key-value pairs are updated according to the vehicle clustering results of the day.

[0106] Based on the above vehicle clustering method, the designated files where vehicles frequently appear are determined in the various vehicle files corresponding to the designated area, and the clustering threshold corresponding to the designated object file is adjusted, which helps to improve the recall rate corresponding to each vehicle file in the vehicle clustering process.

[0107] Based on the same inventive concept, the present application also provides a target object clustering device, such as Figure 2 FIG. 1 is a schematic diagram of a target object clustering device in the present application, and the device includes:

[0108] An acquisition module 21 is used to acquire target object picture frame numbers corresponding to a plurality of target object files, wherein the target object pictures in the plurality of target object files are pictures collected within a target area and within a preset time period;

[0109] A calculation module 22, for calculating the first target object acquisition score corresponding to each target object file in a specified time period according to the number of frames of each target object picture, wherein the specified time period includes a plurality of preset time periods;

[0110] A screening module 23, used to screen out at least one designated object collection score from each first target object collection score, and determine the designated object files corresponding to each designated object collection score;

[0111] The clustering module 24 is used to adjust the clustering thresholds corresponding to the designated object files, and cluster the target object images corresponding to the designated object files according to the adjusted clustering thresholds.

[0112] In a possible design, the acquisition module 21 is specifically used for:

[0113] Acquire multiple target object files corresponding to the target object images collected within the target area and within the specified time period;

[0114] Respectively extracting first trajectory information corresponding to the target object in each target object file within each preset time period;

[0115] According to each piece of first trajectory information, the number of target object picture frames corresponding to each target object file is obtained.

[0116] In a possible design, the acquisition module 21 is further used for:

[0117] According to the collection points corresponding to the target object pictures in each target object file, second trajectory information corresponding to each target object file is obtained;

[0118] The longitude and latitude information of the collection points corresponding to each second track information is converted according to a preset code to obtain third track information;

[0119] The target object images in each of the third track information are deduplicated to obtain the first track information.

[0120] In a possible design, the calculation module 22 is specifically used for:

[0121] According to the number of frames of each target object picture, a plurality of second target object acquisition points corresponding to each target object file are calculated respectively, wherein the same target object file contains pictures acquired in a plurality of preset time periods, and one of the preset time periods corresponds to one second target object acquisition point;

[0122] Setting weight values ​​for each second target object collection score corresponding to the same target object profile, respectively, to obtain third target object collection scores corresponding to a plurality of preset time periods;

[0123] The multiple third target object acquisition scores corresponding to the same target object file are summed up to obtain the first target object acquisition score corresponding to each target object file in the specified time period.

[0124] In a possible design, the screening module 23 is specifically used for:

[0125] Sort the collection scores of the first target objects from largest to smallest;

[0126] The first M first target object acquisition scores among the sorted plurality of first target object acquisition scores are used as designated object acquisition scores;

[0127] The designated object files corresponding to the designated object collection points are determined from the plurality of target object files.

[0128] In a possible design, the clustering module 24 is specifically used for:

[0129] Determining the attenuation value corresponding to each of the designated object files according to the arrangement order of the first target object acquisition points corresponding to each of the designated object files;

[0130] According to the attenuation value, the target object clustering threshold corresponding to the designated object file is adjusted.

[0131] Based on the target object clustering device, the designated object files where the target object frequently appears are determined in each target object file corresponding to the target area, and the clustering threshold corresponding to the designated object file is adjusted, which helps to improve the recall rate corresponding to each target object file during the target object clustering process.

[0132] Based on the same inventive concept, an electronic device is also provided in an embodiment of the present application, and the electronic device can realize the function of the aforementioned target object clustering device, referring to Figure 3 , the electronic device comprises:

[0133] At least one processor 31, and a memory 32 connected to the at least one processor 31. The specific connection medium between the processor 31 and the memory 32 is not limited in the embodiment of the present application. Figure 3 In the example, the processor 31 and the memory 32 are connected via the bus 30. The bus 30 is Figure 3 The connection between other components is shown by bold lines, and is not intended to be limiting. The bus 30 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. Alternatively, the processor 31 can also be called a controller, and there is no limitation on the name.

[0134] In the embodiment of the present application, the memory 32 stores instructions that can be executed by at least one processor 31. The at least one processor 31 can execute the target object clustering method discussed above by executing the instructions stored in the memory 32. The processor 31 can implement Figure 2 The functions of each module in the device shown.

[0135] Among them, the processor 31 is the control center of the device, and can use various interfaces and lines to connect the various parts of the entire control device. By running or executing instructions stored in the memory 32 and calling data stored in the memory 32, the various functions of the device and processing data, the device can be monitored as a whole.

[0136] In a possible design, the processor 31 may include one or more processing units, and the processor 31 may integrate an application processor and a modem processor, wherein the application processor mainly processes an operating system, a user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the modem processor may not be integrated into the processor 31. In some embodiments, the processor 31 and the memory 32 may be implemented on the same chip, and in some embodiments, they may also be implemented separately on separate chips.

[0137] The processor 31 can be a general-purpose processor, such as a central processing unit (CPU), a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, and can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the target object clustering method disclosed in the embodiments of the present application can be directly embodied as a hardware processor to execute, or the hardware and software modules in the processor can be combined and executed.

[0138] The memory 32 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules. The memory 32 may include at least one type of storage medium, such as a flash memory, a hard disk, a multimedia card, a card-type memory, a random access memory (Random Access Memory, RAM), a static random access memory (Static Random Access Memory, SRAM), a programmable read-only memory (Programmable Read Only Memory, PROM), a read-only memory (Read Only Memory, ROM), an electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), a magnetic memory, a disk, an optical disk, etc. The memory 32 is any other medium that can be used to carry or store a desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 32 in the embodiment of the present application can also be a circuit or any other device that can realize a storage function, for storing program instructions and / or data.

[0139] By programming the processor 31, the code corresponding to the target object clustering method described in the above embodiment can be fixed into the chip, so that the chip can execute the target object clustering method when running. Figure 1Steps of the target object clustering method of the embodiment shown. How to design and program the processor 31 is a technique known to those skilled in the art and will not be described in detail here.

[0140] Based on the same inventive concept, an embodiment of the present application further provides a storage medium, which stores computer instructions. When the computer instructions are executed on a computer, the computer executes the target object clustering method discussed above.

[0141] In some possible implementations, various aspects of the target object clustering method provided by the present application may also be implemented in the form of a program product, which includes a program code. When the program product is run on an apparatus, the program code is used to enable the control device to execute the steps of the target object clustering method according to various exemplary implementations of the present application described above in this specification.

[0142] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0143] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0144] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0145] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0146] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A target object clustering method, characterized in that: The method comprises: Obtaining target object picture frame numbers corresponding to a plurality of target object files respectively, wherein the target object pictures in the plurality of target object files are pictures collected within a target area and within a preset time period; Calculate the first target object acquisition score corresponding to each target object file in a specified time period according to the number of frames of each target object picture, wherein the specified time period includes multiple preset time periods, and the first target object acquisition score is used to indicate the frequency of the target objects in the multiple target object files appearing in the target area; Selecting at least one designated object collection score from each first target object collection score, and determining designated object profiles corresponding to each designated object collection score; According to the arrangement order of the first target object acquisition points corresponding to each of the designated object files, the attenuation values ​​corresponding to each of the designated object files are determined, and according to the attenuation values, the clustering thresholds corresponding to the designated object files are adjusted, and the target object images corresponding to each of the designated object files are clustered according to the adjusted clustering thresholds.

2. The method according to claim 1, characterized in that The step of obtaining target object picture frame numbers respectively corresponding to a plurality of target object files comprises: Acquire multiple target object files corresponding to all target object images collected in the target area and within the specified time period; Respectively extracting first trajectory information corresponding to the target object in each target object file within each preset time period; According to each piece of first trajectory information, the number of target object picture frames corresponding to each target object file is obtained.

3. The method according to claim 2, characterized in that The step of respectively extracting first trajectory information corresponding to the target object in each target object file within each preset time period includes: According to the collection points corresponding to the target object pictures in each target object file, second trajectory information corresponding to each target object file is obtained; The longitude and latitude information of the collection points corresponding to each second track information is converted according to a preset code to obtain third track information; The target object images in each of the third track information are deduplicated to obtain the first track information.

4. The method according to claim 1, characterized in that According to the number of frames of each target object picture, the first target object acquisition score corresponding to each target object file in the specified time period is calculated respectively, including: According to the number of frames of each target object picture, a plurality of second target object acquisition points corresponding to each target object file are calculated respectively, wherein the same target object file contains pictures acquired in a plurality of preset time periods, and one of the preset time periods corresponds to one second target object acquisition point; Setting weight values ​​for each second target object collection score corresponding to the same target object profile, respectively, to obtain third target object collection scores corresponding to a plurality of preset time periods; The multiple third target object acquisition scores corresponding to the same target object file are summed up to obtain the first target object acquisition score corresponding to each target object file in the specified time period.

5. The method according to claim 1, characterized in that The step of selecting at least one designated object collection score from each first target object collection score and determining the designated object files corresponding to each designated object collection score includes: Sort the collection scores of the first target objects from largest to smallest; The first M first target object acquisition scores among the sorted plurality of first target object acquisition scores are used as designated object acquisition scores; The designated object files corresponding to the designated object collection points are determined from the plurality of target object files.

6. A target object clustering device, characterized in that: The device comprises: An acquisition module, used to acquire target object picture frame numbers corresponding to a plurality of target object files, wherein the target object pictures in the plurality of target object files are pictures collected within a target area and within a preset time period; A calculation module, used to calculate the first target object acquisition score corresponding to each target object file in a specified time period according to the number of frames of each target object picture, wherein the specified time period includes a plurality of preset time periods, and the first target object acquisition score is used to indicate the frequency of occurrence of the target objects in the plurality of target object files in the target area; A screening module, used to screen out at least one designated object collection score from each first target object collection score, and determine the designated object files corresponding to each designated object collection score; The clustering module is used to determine the attenuation value corresponding to each of the designated object files according to the arrangement order of the first target object collection points corresponding to each of the designated object files; adjust the clustering threshold corresponding to the designated object file according to the attenuation value, and cluster the target object images corresponding to each of the designated object files according to the adjusted clustering threshold.

7. The device according to claim 6, characterized in that The acquisition module is specifically used for: Acquire multiple target object files corresponding to the target object images collected within the target area and within the specified time period; Respectively extracting first trajectory information corresponding to the target object in each target object file within each preset time period; According to each piece of first trajectory information, the number of target object picture frames corresponding to each target object file is obtained.

8. The device according to claim 7, characterized in that The acquisition module is also used for: According to the collection points corresponding to the target object pictures in each target object file, second trajectory information corresponding to each target object file is obtained; The longitude and latitude information of the collection points corresponding to each second track information is converted according to a preset code to obtain third track information; The target object images in each of the third track information are deduplicated to obtain the first track information.

9. The device according to claim 6, characterized in that The calculation module is specifically used for: According to the number of frames of each target object picture, a plurality of second target object acquisition points corresponding to each target object file are calculated respectively, wherein the same target object file contains pictures acquired in a plurality of preset time periods, and one of the preset time periods corresponds to one second target object acquisition point; Setting weight values ​​for each second target object collection score corresponding to the same target object profile, respectively, to obtain third target object collection scores corresponding to a plurality of preset time periods; The multiple third target object acquisition scores corresponding to the same target object file are summed up to obtain the first target object acquisition score corresponding to each target object file in the specified time period.

10. The device according to claim 6, characterized in that The screening module is specifically used for: Sort the collection scores of the first target objects from largest to smallest; The first M first target object acquisition scores among the sorted plurality of first target object acquisition scores are used as designated object acquisition scores; The designated object files corresponding to the designated object collection points are determined from the plurality of target object files.

11. An electronic device, characterized in that: include: Memory, used to store computer programs; A processor, configured to implement the method steps of any one of claims 1 to 5 when executing the computer program stored in the memory.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method steps of any one of claims 1 to 5 are implemented.

Images