Target aggregation method, electronic device and computer-readable storage medium

Abstract

The present application discloses a target archiving method, electronic device, and computer-readable storage medium, which includes: obtaining multiple images to be processed corresponding to different parts of a target; archiving the images to be processed at the same part to obtain a part file corresponding to each part; the part file includes at least one image to be processed, and the image similarity between any image to be processed and at least one other image to be processed exceeds a first threshold; determining the file similarity between part files at the same part, and taking two part files whose file similarity exceeds a second threshold as files to be confirmed; the second threshold is less than the first threshold; based on the spatiotemporal information of the images to be processed in the files to be confirmed, obtaining a mobile identification code matched by the files to be confirmed, and using the mobile identification code to archive the files to be confirmed to update the part file; taking the part files of all parts of the target as the target file of the target. The above scheme can improve the accuracy and recall rate of target archiving.

Description

Technical Field

[0001] The present application relates to the field of big data processing technology, and in particular to a target archiving method, an electronic device, and a computer-readable storage medium. Background Art

[0002] With the advent of the data age, massive amounts of data need to be processed and archived, so that data belonging to the same target object can be classified into the same archive, making the data more standardized. However, in the existing technology, when comparing images of a single target type, images with a similarity exceeding a preset threshold are classified into the same archive. However, due to the influence of objective factors such as the angle of image acquisition, light, and occlusion, images of the same target are easily classified into different archives, resulting in poor accuracy and recall of target archiving. In view of this, how to improve the accuracy and recall of target archiving has become an urgent problem to be solved. Summary of the Invention

[0003] The main technical problem solved by this application is to provide a target clustering method, electronic device and computer-readable storage medium, which can improve the accuracy and recall rate of target clustering.

[0004] In order to solve the above technical problems, the first aspect of the present application provides a target clustering method, comprising: obtaining multiple images to be processed corresponding to different parts of a target; clustering the images to be processed at the same part to obtain a part file corresponding to each part; wherein the part file includes at least one image to be processed, and when multiple images to be processed are included, the image similarity between any one of the images to be processed and at least one other image to be processed exceeds a first threshold; determining the file similarity between the part files at the same part, and taking two part files whose file similarity exceeds a second threshold as files to be confirmed; wherein the second threshold is less than the first threshold; based on the spatiotemporal information of the images to be processed in the files to be confirmed, obtaining a mobile identification code matched by the files to be confirmed, and using the mobile identification code to cluster the files to be confirmed to update the part file; taking the part files corresponding to all parts of the target as the target file of the target.

[0005] To solve the above technical problems, the second aspect of the present application provides an electronic device, which includes: a memory and a processor coupled to each other, wherein the memory stores program data, and the processor calls the program data to execute the method described in the first aspect above.

[0006] In order to solve the above technical problems, the third aspect of the present application provides a computer-readable storage medium on which program data is stored. When the program data is executed by a processor, the method described in the first aspect is implemented.

[0007] The above scheme obtains multiple images to be processed corresponding to different parts of the target, that is, different types of images to be processed corresponding to multiple targets, and based on the image similarity between the images to be processed, the images to be processed at the same part are clustered to obtain a part file corresponding to each part, and each part file includes one or more images to be processed. When the part file includes multiple images to be processed, the image similarity between any image to be processed and at least one other image to be processed exceeds a first threshold, the part files corresponding to the same part are compared to obtain the file similarity between the part files at the same part, and the two part files whose file similarity exceeds a second threshold are used as the part files to be confirmed. Recognize the files, and the second threshold is less than the first threshold, that is, the part files with similarity higher than the second threshold and lower than the first threshold are taken as the files to be confirmed, then the files to be confirmed are multiple files suspected to belong to the same target, which need to be further confirmed. Based on the spatiotemporal information of the images to be processed in the files to be confirmed, the mobile identification codes matched by all the images to be processed in the files to be confirmed are obtained, and the files to be confirmed are clustered twice using the mobile identification codes, so as to update the part files, reduce the probability of multiple files for the same target, improve the recall rate of clustering, merge the part files corresponding to all parts of the target, and obtain the target file of the target, so as to make the target file more comprehensive and accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without inventive efforts. Among them:

[0009] Figure 1 This is a flow chart of an implementation method of the target file aggregation method of the present application;

[0010] Figure 2 This is a flow chart of another embodiment of the target file aggregation method of the present application;

[0011] Figure 3 This is a schematic diagram of an application scenario of an embodiment of the target file aggregation method of the present application;

[0012] Figure 4 This is a schematic structural diagram of an embodiment of the electronic device of the present application;

[0013] Figure 5 It is a structural diagram of an embodiment of a computer-readable storage medium of the present application. DETAILED DESCRIPTION

[0014] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0015] If the technical solution of this application involves personal information, the product that applies the technical solution of this application has clearly informed the personal information processing rules and obtained the individual's voluntary consent before processing personal information. If the technical solution of this application involves sensitive personal information, the product that applies the technical solution of this application has obtained the individual's separate consent before processing sensitive personal information, and at the same time meets the "explicit consent" requirement. For example, on personal information collection devices such as cameras, a clear and prominent sign is set to inform that the personal information collection scope has been entered and personal information will be collected. If the individual voluntarily enters the collection scope, it is deemed that they agree to the collection of their personal information; or on the personal information processing device, when the personal information processing rules are notified by obvious signs / information, the individual's authorization is obtained through pop-up information or by asking the individual to upload their personal information; among which, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the type of personal information processed.

[0016] The terms "system" and "network" are often used interchangeably in this document. The term "and / or" is simply a description of an association between related objects, indicating that three possible relationships exist. For example, "A and / or B" can mean: A exists alone, A and B exist simultaneously, or B exists alone. Furthermore, the character " / " generally indicates that the related objects are in an "or" relationship. Furthermore, "multiple" in this document means two or more than two.

[0017] For ease of understanding, the terms involved in the embodiments of the present application are explained below, where "aggregate file" means generating archival information of a target, "multi-file" means that the actual archival subject of multiple archives is the same target, and "recall rate" means that in an archive, all the images to be processed of the target need to be aggregated as much as possible. For example, if a target object has 100 images to be processed, but there are only 80 in the archive, the recall rate is 80 / 100=80%.

[0018] The target clustering method provided in this application is used to cluster the images to be processed corresponding to the target, and the target includes a person or an object. The corresponding execution subject of the target clustering method provided in this application is a processor that can call images or videos.

[0019] See also Figure 1 , Figure 1: is a flow chart of an embodiment of the target file aggregation method of the present application, the method comprising:

[0020] S101: Obtaining multiple images to be processed corresponding to different parts of the target.

[0021] Specifically, multiple images to be processed corresponding to different parts of the target are obtained, that is, different types of images to be processed corresponding to the multiple targets are obtained.

[0022] In one application, an overall image of multiple targets is obtained, and the overall image and images of at least a portion of a specified portion extracted from the overall image are used as images to be processed, thereby obtaining the overall image and segmenting the overall image to obtain the image to be processed.

[0023] In another application mode, images to be processed uploaded by multiple data sources are obtained, wherein each data source includes images to be processed corresponding to different parts of the target, thereby directly obtaining the images to be processed.

[0024] S102: Aggregate the images to be processed at the same part to obtain a part file corresponding to each part, wherein the part file includes at least one image to be processed. When the part file includes multiple images to be processed, the image similarity between any image to be processed and at least one other image to be processed exceeds a first threshold.

[0025] Specifically, based on the image similarity between the images to be processed, the images to be processed at the same part are clustered to obtain a part file corresponding to each part. Each part file includes one or more images to be processed. When the part file includes multiple images to be processed, the image similarity between any image to be processed and at least one other image to be processed exceeds a first threshold.

[0026] In one application method, for the images to be processed corresponding to each part, the image similarity between the images to be processed is determined, and the two images to be processed with the highest image similarity are first clustered into a set, and the images to be processed whose image similarity with any image to be processed in the set exceeds a first threshold value are clustered into the corresponding set. When the image similarity between the image to be processed and the images to be processed in multiple sets exceeds the first threshold value, the corresponding image to be processed is clustered into the set with the highest image similarity, thereby obtaining a part file including multiple images to be processed, and the other images to be processed outside the set are separately regarded as a part file.

[0027] In another application method, for the images to be processed corresponding to each part, the image similarity between the images to be processed is determined, and the two images to be processed with the highest image similarity are first clustered into a set. The images to be processed whose average image similarity with all the images to be processed in the set exceeds a first threshold are clustered into the corresponding set to obtain a part file including multiple images to be processed, and the other images to be processed outside the set are treated as a separate part file.

[0028] S103: Determine the file similarity between the part files at the same part, and take two part files whose file similarity exceeds a second threshold as files to be confirmed, wherein the second threshold is less than the first threshold.

[0029] Specifically, the part files corresponding to the same part are compared to obtain the file similarity between the part files at the same part, and the two part files whose file similarity exceeds the second threshold are taken as files to be confirmed, and the second threshold is less than the first threshold, that is, the part files with a similarity higher than the second threshold and lower than the first threshold are taken as files to be confirmed. The files to be confirmed are multiple files suspected to belong to the same target and need to be further confirmed.

[0030] In one application, the image similarity between the to-be-processed images in two part files of the same part is obtained, and the image similarities between the to-be-processed images in the two part files of the same part are averaged to obtain the file similarity between the two part files.

[0031] In another application method, a representative image is selected from two part files of the same part respectively, wherein, when the part file includes multiple images to be processed, the representative image is the image with the highest average image similarity with other images to be processed; when the part file includes one image to be processed, the representative image is the corresponding image to be processed, and based on the image similarity between the representative images, the file similarity between the two part files is determined.

[0032] S104: Based on the spatiotemporal information of the image to be processed in the file to be confirmed, a mobile identification code matching the file to be confirmed is obtained, and the file to be confirmed is aggregated using the mobile identification code to update the part file.

[0033] Specifically, based on the spatiotemporal information of the images to be processed in the files to be confirmed, the mobile identification codes matching all the images to be processed in the files to be confirmed are obtained, and the files to be confirmed are secondary clustered using the mobile identification codes, thereby updating the part files, reducing the probability of multiple files for the same target, and improving the recall rate of clustered files.

[0034] In one application method, the spatiotemporal information of the images to be processed in the to-be-confirmed file is obtained, and the mobile identification codes matched by all the images to be processed in the to-be-confirmed file are obtained based on the spatiotemporal information. The mobile identification codes matched by the two to-be-confirmed files are compared one by one to obtain the matching rate of the mobile identification codes. The to-be-confirmed files are clustered based on the matching rate, and the to-be-confirmed files with a matching rate exceeding a third threshold are merged, thereby updating the part files, reducing the probability of multiple files for the same target, and improving the accuracy and recall rate of target clustering.

[0035] In another application method, in addition to the file similarity, the files to be confirmed also include the initial confidence, the spatiotemporal information of the images to be processed in the files to be confirmed is obtained, and the mobile identification codes matched by all the images to be processed in the files to be confirmed are obtained based on the spatiotemporal information. The mobile identification codes matched by the two files to be confirmed are compared one by one to obtain the matching rate of the mobile identification codes, and the matching rate is converted into a corrected confidence, wherein the corrected confidence is positively correlated with the matching rate. The target confidence is obtained based on the sum of the initial confidence and the corrected confidence, and the files to be confirmed whose target confidence exceeds the confidence threshold are merged, thereby updating the part files, reducing the probability of multiple files for the same target, and improving the accuracy and recall rate of target clustering.

[0036] S105: The part files corresponding to all parts of the target are used as the target files of the target.

[0037] Specifically, the part files corresponding to all parts of the target are merged to obtain the target file of the target, so that the target file is more comprehensive and accurate.

[0038] In one application, there is an association relationship between the parts of the target. Based on the part file corresponding to one part of the target, the part files of other parts corresponding to the target are obtained according to the association relationship, and the part files corresponding to all parts of the target are combined to form the target file of the target.

[0039] In another application, the target part includes the entire region and partial region of the target, and the part file matching the entire region corresponding to the target and the part file matching the partial region corresponding to the target are merged to obtain the target file of the target.

[0040] In one application scenario, the parts corresponding to the target include at least the head and torso of the target, and the torso corresponds to the back and the front, wherein the first threshold corresponding to the head of the target when the file is gathered is greater than the first threshold corresponding to the torso when the file is gathered, and the second threshold corresponding to the head of the target when the file is gathered is greater than the second threshold corresponding to the torso when the file is gathered; the part files corresponding to all parts of the target are used as the target file of the target, including: based on the association relationship between the head and torso corresponding to the target, obtaining the part files corresponding to the head, the back of the torso and the front of the torso corresponding to the target as the target file of the target.

[0041] Specifically, the parts of the target include the head and torso of the target, and the torso has at least a front and a back, thereby enriching the dimension of the image, and the first threshold corresponding to the head of the target during clustering is greater than the first threshold corresponding to the torso during clustering, so that the clustering threshold corresponding to the area with rich feature information is greater than the clustering threshold corresponding to the area with less feature information, thereby improving the accuracy of clustering.

[0042] Furthermore, when determining the target file corresponding to the target, the part file corresponding to the target's head and the part files of the back and front of the target's body are collectively used as the target file of the target, so that the target file includes at least one part file corresponding to the head and part files of two angles on the body, thereby improving the accuracy of the target file.

[0043] In a specific application scenario, the values ​​of image similarity and file similarity are between 0-1, the first threshold corresponding to the target's head when clustered is greater than the first threshold corresponding to the target's body when clustered, and the second threshold is the first threshold of the corresponding part minus a fixed difference. Therefore, the second threshold corresponding to the part file corresponding to the target's head when clustered is always greater than the second threshold corresponding to the part file corresponding to the target's body when clustered. The above values ​​are only examples of the first threshold and the second threshold corresponding to different parts, and the first threshold and the second threshold can be customized, and this application does not impose specific restrictions on this.

[0044] The above scheme obtains multiple images to be processed corresponding to different parts of the target, that is, different types of images to be processed corresponding to multiple targets, and based on the image similarity between the images to be processed, the images to be processed at the same part are clustered to obtain a part file corresponding to each part, and each part file includes one or more images to be processed. When the part file includes multiple images to be processed, the image similarity between any image to be processed and at least one other image to be processed exceeds a first threshold, the part files corresponding to the same part are compared to obtain the file similarity between the part files at the same part, and the two part files whose file similarity exceeds a second threshold are used as the part files to be confirmed. Recognize the files, and the second threshold is less than the first threshold, that is, the part files with similarity higher than the second threshold and lower than the first threshold are taken as the files to be confirmed, then the files to be confirmed are multiple files suspected to belong to the same target, which need to be further confirmed. Based on the spatiotemporal information of the images to be processed in the files to be confirmed, the mobile identification codes matched by all the images to be processed in the files to be confirmed are obtained, and the files to be confirmed are clustered twice using the mobile identification codes, so as to update the part files, reduce the probability of multiple files for the same target, improve the recall rate of clustering, merge the part files corresponding to all parts of the target, and obtain the target file of the target, so as to make the target file more comprehensive and accurate.

[0045] See also Figure 2 , Figure 2 : is a flow chart of another embodiment of the target file aggregation method of the present application, the method comprising:

[0046] S201: Acquire data to be processed sent by at least one data source, wherein the data to be processed includes a signature certificate matching the data source and an initial image from the data source, wherein the initial image corresponds to different parts of multiple targets.

[0047] Specifically, see Figure 3 , Figure 3 This is a schematic diagram of an application scenario of an embodiment of the target archiving method of the present application. The target archiving method is implemented based on a target archiving system, wherein the target archiving system includes four units: data acquisition, data storage, data modeling, and data application. The specific modules corresponding to each unit are as follows: Figure 3 shown.

[0048] Furthermore, data to be processed sent by one or more data sources is obtained, wherein the data to be processed includes a signature certificate matching the data source and an initial image from the data source, and the initial image includes different parts of different targets, thereby improving the scalability and adaptability of the target aggregation method.

[0049] S202: Verify the signature certificate included in the data to be processed to obtain the verified data to be processed, perform data management on the initial image included in the verified data to be processed, and obtain multiple images to be processed corresponding to different parts of the target.

[0050] Specifically, the signature certificate included in the data to be processed is verified, and the data to be processed that passes the verification is determined, thereby improving the security of data processing.

[0051] Furthermore, data governance is performed on the initial images included in the verified data to be processed, resulting in multiple images corresponding to different target parts. Data governance includes streamlining business processes, planning data resources, and performing data deduplication, desensitization, conversion, correlation, and anomaly removal. This ensures data planning, management, storage, and application, improving the accuracy of the images to be processed.

[0052] It should be noted that the signature certificate is updated based on a preset period. The update process of the signature certificate includes: using a preset interface to obtain the request information sent by the data source; wherein the request information includes a request key and a certificate address, and the request key is generated based on a symmetric encryption algorithm; in response to successful verification of the request key, the signature certificate of the data source is obtained based on the certificate address, and the obtained signature certificate is marked as a trusted certificate.

[0053] Specifically, multiple data sources send request information through a preset interface, thereby obtaining the request information using the preset interface. The request information includes a request key and a certificate address, and the request key is generated based on a symmetric encryption algorithm. The request key is verified after obtaining the request key. When the request key verification is successful, the certificate address matching the successfully verified request key is obtained, and the signature certificate of the corresponding data source is obtained using the certificate address. The obtained signature certificate is marked as a trusted certificate, thereby improving the security and privacy of data processing through periodic verification.

[0054] In one application scenario, the preset interface is a Restful interface. Multiple data sources complete data access and upload request information through the Restful interface. The request information includes a fixed request key, which is encrypted according to a symmetric encryption algorithm. After successful verification, the Restful interface response result is parsed to obtain the certificate address, the signed certificate is downloaded, and stored in the trusted area, so that the signed certificate in the trusted area is marked as a trusted certificate.

[0055] Optionally, if there is no Restful interface authentication, trusted data authentication is performed through the third-party software development tool (Software Development Kit, SDK) interface, and the signature certificate is downloaded for storage. After the signature certificate is saved, multiple data sources upload data in a unified access data model in the form of a signaling stream or data stream carrying the signature certificate, so that the uploaded data can be trusted and processed within a preset period, wherein the signature certificate includes the data source, data type, data time and data protocol.

[0056] Please refer again to Figure 3 The data acquisition unit is used to verify the label certificate and obtain reliable data, thereby improving the security of the data, and then input the data to be processed into the data storage unit for storage. The relevant data storage method can be any method in the existing technology, and this application does not elaborate on this. The data processing unit is used to perform data management on the data, obtain the image to be processed, and improve the accuracy of the image to be processed.

[0057] S203: Aggregate the images to be processed at the same part to obtain a part file corresponding to each part, wherein the part file includes at least one image to be processed. When multiple images to be processed are included, the image similarity between any image to be processed and at least one other image to be processed exceeds a first threshold.

[0058] Specifically, the images to be processed of the same part are compared, and based on the image similarity between the images to be processed, the images to be processed at the same part are clustered to obtain a part file corresponding to each part.

[0059] In one application, image features corresponding to the image to be processed are extracted, and based on the image features, the image similarity between any two images to be processed at the same part is determined; wherein the image features are related to the attributes and characteristic values ​​of the image to be processed; the images to be processed whose image similarity exceeds a first threshold are clustered into the same set, and the set and the images to be processed outside the set are respectively regarded as a file to obtain a part file corresponding to each part.

[0060] Specifically, image features corresponding to the image to be processed are extracted, wherein the image features are related to the attributes and feature values ​​of the image to be processed, and the attributes and feature values ​​of the image to be processed are related to the type of the part, thereby improving the accuracy of the image features.

[0061] Furthermore, the image features between the images to be processed are compared to determine the image similarity between any two images to be processed at the same part, thereby improving the accuracy of the image similarity, clustering the images to be processed whose image similarity exceeds a first threshold into the same set, and treating the set as a part file including multiple images to be processed, and treating the images to be processed outside the set as a part file including only one image to be processed, thereby obtaining a part file corresponding to each part.

[0062] In one application scenario, the target is a person, and the attributes corresponding to the images to be processed of different parts of the target include the target's coordinates and age, and the feature values ​​include the target's age and gender. After obtaining the image features, the similarity between the image features is compared based on the cosine algorithm. The higher the cosine value, the higher the similarity.

[0063] S204: Determine the file similarity between the part files at the same part, and take the two part files whose file similarity exceeds a second threshold as files to be confirmed, wherein the second threshold is less than the first threshold.

[0064] Specifically, the part files corresponding to the same part are compared to obtain the file similarity between the part files at the same part, and the two part files whose file similarity exceeds the second threshold are taken as files to be confirmed, and the second threshold is less than the first threshold, so as to obtain a part file with higher similarity but the similarity does not exceed the first threshold as the file to be confirmed.

[0065] In one application method, the file similarity between part files at the same part is determined, and two part files whose file similarity exceeds a second threshold are used as files to be confirmed, including performing the following steps for the part files at the same part: obtaining the number of images to be processed in the part file, and determining the file weight corresponding to each part file based on the number of images; wherein the file weight is positively correlated with the number of images; obtaining two part files whose sum of file weights is less than the weight threshold, and using the average value of the image similarity between the images to be processed in the corresponding two part files as the file similarity between the part files; and using the two part files whose file similarity exceeds the second threshold as files to be confirmed.

[0066] Specifically, the total number of images to be processed in the part file is taken as the number of images, and a corresponding file weight is set for each part file according to the number of images. The file weight is positively correlated with the number of images, that is, the larger the number of images, the higher the file weight.

[0067] Furthermore, the file weights of the two part files are added together to obtain two part files whose sum of the file weights is less than the weight threshold. The images to be processed in one part file are compared with the images to be processed in the other part file to determine the image similarity. The average value of the image similarities between the images to be processed in the two corresponding part files is used as the file similarity between the part files. By setting the file weights, it is convenient to merge a part file with a smaller number of images to be processed with a part file with a larger number of images to be processed, or to merge two part files with a smaller number, and reduce the number of image comparisons when determining the file similarity between the part files to improve the file aggregation efficiency.

[0068] It is understandable that the two part files whose file similarity exceeds the second threshold are taken as files to be confirmed, and further confirmation is performed to determine whether they need to be merged.

[0069] In one application scenario, the file weight is set to correspond to the order of magnitude of the number of images corresponding to all part files. The order of magnitude corresponding to the number of images with the largest value is used as a reference value. For example, if the number of images with the largest value is 80, then the order of magnitude of the number of images is in the tens, and the reference value is 10, so the corresponding file weight is 8. At this time, the file weight corresponding to the part file that only includes one image to be processed is 0.1. The weight threshold is the product of the reference value and the preset ratio, where the preset ratio is between 0 and 1. For example, if the preset ratio is 0.5, the weight threshold is 5. Of course, the preset ratio can also be customized in other application scenarios, and this application does not impose specific restrictions on this.

[0070] Optionally, the parts corresponding to the target include at least the head and torso of the target, and the torso corresponds to the back and the front, wherein the first threshold corresponding to the head of the target when the gear is gathered is greater than the first threshold corresponding to the torso when the gear is gathered, and the second threshold corresponding to the head of the target when the gear is gathered is greater than the second threshold corresponding to the torso when the gear is gathered.

[0071] S205: Based on the spatiotemporal information of the image to be processed in the file to be confirmed, a mobile identification code matching the file to be confirmed is obtained, and the file to be confirmed is aggregated using the mobile identification code to update the part file.

[0072] Specifically, the spatiotemporal information includes the acquisition time and location corresponding to the image to be processed. Based on the acquisition time and location corresponding to the image to be processed in the file to be confirmed, the mobile identification code matching the file to be confirmed is obtained, and the mobile identification code is used to aggregate the file to be confirmed to update the part file.

[0073] In one application method, based on the acquisition time and acquisition location corresponding to the image to be processed in the file to be confirmed, a mobile identification code matching the acquisition time and acquisition location is obtained to obtain the mobile identification code matching the file to be confirmed; the mobile identification codes matching each of the files to be confirmed are compared to obtain the matching rate between the files to be confirmed; based on the matching rate, the part files are aggregated to update the part files.

[0074] Specifically, each target has a unique mobile identification code, and the mobile identification code matches the target's location. When the target appears at a certain location at a certain time, the mobile identification code corresponding to the target can be obtained. Based on the acquisition time and acquisition location corresponding to the image to be processed in the file to be confirmed, the mobile identification code matching the acquisition time and acquisition location is obtained, and the mobile identification code matching the file to be confirmed is obtained. By comparing the mobile identification codes matching each of the files to be confirmed, the matching rate between the part files is obtained, thereby improving the accuracy of the matching rate.

[0075] Furthermore, the matching rate is compared with the matching threshold, and the part files that reach the matching threshold are merged to reduce the probability of multiple files for the same target. The part files that do not reach the matching threshold are still stored separately, thereby improving the recall rate of the aggregated files.

[0076] In one application scenario, the mobile identification codes that match each of the files to be confirmed are compared to obtain a matching rate between the files to be confirmed, including: obtaining trajectory information or location information corresponding to the files to be confirmed based on the mobile identification codes that match each of the files to be confirmed; wherein, when the file to be confirmed includes multiple images to be processed, the corresponding file to be confirmed has corresponding trajectory information, and when the file to be confirmed includes one image to be processed, the corresponding file to be confirmed has corresponding location information; based on the trajectory information or location information corresponding to each of the files to be confirmed, the matching rate between the files to be confirmed is determined.

[0077] Specifically, the trajectory information or location information of the target corresponding to the image to be processed in the file to be confirmed is restored based on the mobile identification code. When the file to be confirmed includes multiple images to be processed, the corresponding file to be confirmed has trajectory information. When the file to be confirmed includes one image to be processed, the corresponding file to be confirmed has location information.

[0078] Furthermore, the trajectory information or location information between the files to be confirmed is compared, so as to determine the matching rate between the files to be confirmed based on the trajectory information or location information converted from the mobile identification code, thereby improving the accuracy of the matching rate compared with the image comparison-based method.

[0079] It should be noted that when two files to be confirmed correspond to two trajectory information, the trajectory information is predicted and compared, and the overlap rate between the trajectory information is determined to obtain the matching rate between the files to be confirmed. When the files to be confirmed correspond to trajectory information and location information, the trajectory information is predicted, and it is determined whether the location information is located on the trajectory information or the predicted trajectory information. The overlap rate between the location information and the trajectory information is determined to obtain the matching rate between the files to be confirmed. When the files to be confirmed correspond to two location information, the corresponding distance and speed between the location information are calculated, the contradiction probability between the location information is determined, and the matching rate is set between the files to be confirmed based on the contradiction probability.

[0080] S206: The part files corresponding to all parts of the target are used as the target files of the target.

[0081] Specifically, the part files corresponding to all parts of the target are merged to obtain the target file of the target.

[0082] Furthermore, when the parts corresponding to the target include at least the head and torso of the target, and the torso has a back and a front, based on the association relationship between the head and torso corresponding to the target, the part files corresponding to the head, back of the torso and front of the torso corresponding to the target are obtained as the target files of the target.

[0083] Specifically, based on the association between the head and torso corresponding to the target, a part file corresponding to the head of the target and two part files corresponding to the torso of the target are combined to form a target file of the target, so that the target file is more comprehensive and accurate.

[0084] Please refer again to Figure 3 ,The data modeling unit is used to finally cluster the images to be processed into ,target files corresponding to the targets, and send the target files to the ,data application unit, so that the target files can be applied in different ,application scenarios.

[0085] See also Figure 4 , Figure 4 This is a structural diagram of an embodiment of an electronic device of the present application. The electronic device 40 includes a memory 401 and a processor 402 coupled to each other, wherein the memory 401 stores program data (not shown in the figure), and the processor 402 calls the program data to implement the method in any of the above embodiments. For an explanation of the relevant content, please refer to the detailed description of the above method embodiments, which will not be repeated here.

[0086] See also Figure 5 , Figure 5 This is a structural diagram of an embodiment of a computer-readable storage medium of the present application. The computer-readable storage medium 50 stores program data 500. When the program data 500 is executed by the processor, the method in any of the above embodiments is implemented. For an explanation of the relevant content, please refer to the detailed description of the above method embodiments, which will not be repeated here.

[0087] It should be noted that the units described as separate components may or may not be physically separate, and 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 may be selected according to actual needs to achieve the purpose of this embodiment.

[0088] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0089] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, 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, which is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of each embodiment method of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0090] The above description is only an implementation method of the present application and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the description and drawings of this application, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A target aggregation method, characterized in that: The method comprises: Obtaining multiple images to be processed corresponding to different parts of the target; Aggregating the images to be processed at the same part to obtain a part file corresponding to each part; wherein the part file includes at least one image to be processed, and when the part file includes multiple images to be processed, the image similarity between any one of the images to be processed and at least one other image to be processed exceeds a first threshold; Determine the file similarity between the part files at the same part, and take the two part files whose file similarity exceeds a second threshold as files to be confirmed; wherein the second threshold is less than the first threshold; Based on the spatiotemporal information of the image to be processed in the to-be-confirmed file, obtaining a mobile identification code matched with the to-be-confirmed file, and using the mobile identification code to aggregate the to-be-confirmed file to update the part file; The part files corresponding to all parts of the target are used as the target files of the target.

2. The target aggregation method according to claim 1, characterized in that: The step of aggregating the images to be processed at the same part to obtain a part file corresponding to each part includes: Extracting image features corresponding to the image to be processed, and determining image similarity between any two images to be processed at the same location based on the image features; wherein the image features are related to attributes and feature values ​​of the image to be processed; The images to be processed whose image similarity exceeds the first threshold are clustered into the same set, and the set and the images to be processed outside the set are respectively regarded as a file to obtain a part file corresponding to each part.

3. The target aggregation method according to claim 2, characterized in that: Determining the file similarity between the part files at the same part, and taking the two part files whose file similarity exceeds a second threshold as files to be confirmed, includes performing the following steps on the part files at the same part: Obtaining the number of the to-be-processed images in the part file, and determining a file weight corresponding to each of the part files based on the number of images; wherein the file weight is positively correlated with the number of images; Obtaining two part files whose sum of file weights is less than a weight threshold, and taking an average value of image similarities between the to-be-processed images in the corresponding two part files as the file similarity between the part files; The two part files whose file similarity exceeds the second threshold are used as files to be confirmed.

4. The target aggregation method according to claim 1, characterized in that: The spatiotemporal information includes the acquisition time and location corresponding to the image to be processed. The acquiring of the mobile identification code matched with the file to be confirmed based on the spatiotemporal information of the image to be processed in the file to be confirmed, and archiving the file to be confirmed using the mobile identification code to update the part file includes: Based on the acquisition time and acquisition location corresponding to the image to be processed in the to-be-confirmed file, obtaining a mobile identification code that matches the acquisition time and acquisition location, thereby obtaining a mobile identification code that matches the to-be-confirmed file; Comparing the mobile identification codes that match the files to be confirmed to obtain a matching rate between the files to be confirmed; Based on the matching rate, the part file is aggregated to update the part file.

5. The target aggregation method according to claim 4, characterized in that: The comparing the mobile identification codes of the files to be confirmed to obtain a matching rate between the files to be confirmed includes: Based on the mobile identification codes matched to the respective files to be confirmed, the trajectory information or location information corresponding to the files to be confirmed is obtained; wherein, when the files to be confirmed include multiple images to be processed, the corresponding files to be confirmed have the trajectory information; and when the files to be confirmed include one image to be processed, the corresponding files to be confirmed have the location information; Based on the trajectory information or the location information corresponding to each of the files to be confirmed, a matching rate between the files to be confirmed is determined.

6. The target aggregation method according to claim 1, characterized in that: The step of obtaining a plurality of images to be processed corresponding to different parts of the target includes: Obtaining data to be processed sent by at least one data source; wherein the data to be processed includes a signature certificate matching the data source and an initial image from the data source; wherein the initial image corresponds to different parts of multiple targets; The signature certificate included in the data to be processed is verified to obtain the verified data to be processed, and data management is performed on the initial image included in the verified data to be processed to obtain multiple images to be processed corresponding to different parts of the target.

7. The target aggregation method according to claim 6, characterized in that: The signature certificate is updated based on a preset period, and the updating process of the signature certificate includes: Obtaining the request information sent by the data source using a preset interface; wherein the request information includes a request key and a certificate address, and the request key is generated based on a symmetric encryption algorithm; In response to the request key verification being successful, the signature certificate of the data source is obtained based on the certificate address, and the obtained signature certificate is marked as a trusted certificate.

8. The target aggregation method according to any one of claims 1 to 7, characterized in that: The target corresponding parts include at least the target's head and torso, and the torso has a back and a front, wherein a first threshold corresponding to the target's head when the target is in the focus position is greater than a first threshold corresponding to the torso when the target is in the focus position, and a second threshold corresponding to the target's head when the target is in the focus position is greater than a second threshold corresponding to the torso when the target is in the focus position; The step of using the part files corresponding to all parts of the target as the target file of the target includes: Based on the association relationship between the head and torso corresponding to the target, the part files corresponding to the head, the back of the torso and the front of the torso corresponding to the target are obtained as the target file of the target.

9. An electronic device, characterized in that: include: A memory and a processor coupled to each other, wherein the memory stores program data, and the processor calls the program data to execute the method according to any one of claims 1 to 8.

10. A computer-readable storage medium having program data stored thereon, characterized in that: When the program data is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images