A target judgment processing method and system based on multi-image retrieval data fusion

By setting similarity thresholds in multi-graph retrieval and performing data fusion processing, the problems of resource waste and inefficiency in multi-graph retrieval are solved, more efficient resource use and deduplication processing are achieved, and the efficiency of multi-graph retrieval is improved.

CN114020946BActive Publication Date: 2024-12-06CHONGQING UNISINSIGHT TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202111355525.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-11-16
Publication Date
2024-12-06
Estimated Expiration
2041-11-16

AI Technical Summary

Technical Problem

The existing multi-picture retrieval technology has problems such as high time complexity and waste of resources. Especially when judging whether multiple images are the same target, frequent access to the database is required, resulting in inefficient query and excessive resource consumption.

Method used

Using a processing method based on multi-picture search data fusion, the feature similarity and data similarity thresholds are set in advance, the image similarity is judged and the fusion process is performed, unnecessary image query and download are reduced, and resource usage is optimized through deduplication processing.

Benefits of technology

It effectively reduces the waste of resources, especially bandwidth and memory consumption, reduces the cost of deduplication, and improves the efficiency of multi-graph retrieval.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114020946B_ABST
    Figure CN114020946B_ABST
Patent Text Reader

Abstract

The present invention provides a target judgment processing method and system based on multi-image retrieval data fusion, the method comprising: presetting a first comparison threshold for image feature similarity comparison and a second comparison threshold for data result similarity comparison; collecting image information, and performing multi-image retrieval on the image to be processed according to a retrieval request; when the similarity comparison result of the image to be processed reaches the first comparison threshold, determining that the initial images are similar, and performing a first fusion process; when the similarity comparison result of the image to be processed does not reach the first comparison threshold, determining that the initial images are not similar, and if the data duplication reaches the second comparison threshold, determining that the processing objects are the same target, and performing a second fusion process; the present invention saves protected bandwidth and memory and CPU resources, and greatly reduces the cost of deduplication.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of computer applications, and in particular to a target judgment processing method and system based on multi-graph retrieval data fusion. Background Art

[0002] As the data based on parsed algorithms increases, the related application scenarios also increase. Therefore, the need to retrieve multiple images together has also arisen. At present, for multi-image retrieval, the commonly used method is to search through multiple queries in the database. However, this method requires multiple visits to the database, so it has a high time complexity, and frequent visits to the database are prone to low query efficiency, resulting in excessive consumption of database resources. In practical applications, in order to determine whether the data is the same through multi-image retrieval, the existing method will download a lot of useless image data. On the one hand, due to the existence of a large amount of duplicate data, the cost of deduplication is high. On the other hand, it will occupy a large amount of network resources and system resources, resulting in unnecessary waste of resources. Summary of the invention

[0003] In view of the above-mentioned shortcomings of the prior art, the present invention provides a target judgment processing method and system based on multi-image retrieval data fusion to solve the above-mentioned technical problems.

[0004] The target judgment processing method based on multi-image retrieval data fusion provided by the present invention includes:

[0005] Presetting a first comparison threshold for image feature similarity comparison and a second comparison threshold for data result similarity comparison;

[0006] Collect image information, perform multi-image retrieval on the image to be processed according to the retrieval request, and obtain the similarity of the image to be processed;

[0007] When the similarity comparison result of the image to be processed reaches the first comparison threshold, the initial images are determined to be similar, and a first fusion process is performed;

[0008] When the similarity comparison result of the image to be processed does not reach the first comparison threshold, it is determined that the initial image is not similar, an image query is performed on the image to be processed, and the number of repeated images is obtained based on the address information of the image data in the query result. If the data repetition reaches the second comparison threshold, it is determined that the processing objects are the same target, and a second fusion process is performed.

[0009] In one embodiment of the present invention, the first fusion process includes:

[0010] Determining whether to perform first data fusion;

[0011] If not, the to-be-processed images are respectively subjected to image query and image download;

[0012] If fusion occurs, the images to be processed are deduplicated before being downloaded;

[0013] The second fusion process includes:

[0014] determining whether to perform second data fusion;

[0015] If not, download the image to be processed;

[0016] If fused, the images to be processed will be deduplicated before being downloaded.

[0017] In one embodiment of the present invention, when the similarity comparison result of the image to be processed does not reach the first comparison threshold, determining that the initial image is not similar includes:

[0018] Performing image query on the image to be processed;

[0019] The address information of the image data in the query result is stored in a set whose elements are not repeated, and the number of repeated images is obtained according to the length of the set;

[0020] The data repetition degree is acquired according to the number of repeated images, and the second fusion process is performed when the data repetition degree reaches a second comparison threshold.

[0021] In one embodiment of the present invention, the address information includes a uniform resource locator,

[0022] Performing multiple image queries on the image to be processed;

[0023] After obtaining the query results, store the uniform resource locator of the image in each piece of data into a set with no duplicate elements;

[0024] According to the length of the set and the total number of actual search requests, the number of repeated images is obtained, and then the data duplication degree is obtained.

[0025] In one embodiment of the present invention, the first fusion process further includes:

[0026] If not, the query results of each image are displayed separately, and when loading image data, the image is loaded according to the downloaded image data of the set;

[0027] If fusion is performed, multiple image queries are performed on the image to be processed, and the images are sorted according to their similarity. After all data are deduplicated, the optimal value is taken for feedback.

[0028] In one embodiment of the present invention, duplicate data is obtained according to the difference between the total number of data in the image query and the set length;

[0029] When fusion is selected in the first fusion process, the pictures are sorted according to the similarity of image features. After all data are deduplicated, the top K data are fed back. If there are less than K data, all the deduplicated data are fed back.

[0030] The data duplication in the second fusion process is obtained by the ratio of the repeated data to the query data.

[0031] In one embodiment of the present invention, a third comparison threshold for indicating identical data is preset, and the third comparison threshold is greater than the first comparison threshold;

[0032] When the similarity comparison result of the images to be processed reaches a third comparison threshold, it is determined that the processing objects are the same target, and the second fusion processing is performed.

[0033] The present invention also provides a target judgment processing system based on multi-image retrieval data fusion, comprising:

[0034] A configuration module, used to pre-set a first comparison threshold for image feature similarity comparison and a second comparison threshold for data result similarity comparison;

[0035] Image acquisition module, used to collect image information,

[0036] A retrieval module is used to perform multi-image retrieval on the image to be processed according to the retrieval request and obtain the similarity of the image to be processed;

[0037] A processing module, configured to determine that the initial images are similar and perform a first fusion process when the similarity comparison result of the image to be processed reaches the first comparison threshold;

[0038] When the similarity comparison result of the image to be processed does not reach the first comparison threshold, it is determined that the initial image is not similar, an image query is performed on the image to be processed, and the number of repeated images is obtained based on the address information of the image data in the query result. If the data repetition reaches the second comparison threshold, it is determined that the processing objects are the same target, and a second fusion process is performed.

[0039] Beneficial effects of the present invention: The target judgment processing method and system based on multi-image retrieval data fusion in the present invention first determines whether the data is fused. If fused, the image addresses are sorted and deduplicated for download. If not fused, the images are downloaded through all the image addresses of the elements in all the set collections fed back by the query request. The amount of duplicate data is determined by the length of the set collection and the actual returned data, which saves protected bandwidth, memory and CPU resources, and greatly reduces the cost of deduplication. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a flowchart of a target judgment processing method based on multi-graph retrieval data fusion in an embodiment of the present invention.

[0041] Figure 2 It is a flow chart of Embodiment 1 of a target judgment processing method based on multi-graph retrieval data fusion in an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0043] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0044] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0045] like Figure 1 As shown, the target judgment processing method based on multi-image retrieval data fusion in this embodiment includes:

[0046] S1. Preset a first comparison threshold for image feature similarity comparison and a second comparison threshold for data result similarity comparison;

[0047] S2. Collect image information, and perform multi-image retrieval on the image to be processed according to the retrieval request to obtain the similarity of the image to be processed;

[0048] S3. When the similarity comparison result of the image to be processed reaches the first comparison threshold, the initial image is determined to be similar, and a first fusion process is performed;

[0049] S4. When the similarity comparison result of the image to be processed does not reach the first comparison threshold, it is determined that the initial image is not similar, an image query is performed on the image to be processed, and the number of repeated images is obtained based on the address information of the image data in the query result. If the data repetition reaches the second comparison threshold, it is determined that the processing objects are the same target, and a second fusion process is performed.

[0050] In step S1 of the present embodiment, firstly, a first comparison threshold for image feature similarity comparison and a second comparison threshold for data result similarity comparison are configured, as well as a third comparison threshold for indicating identical data. If the similarity threshold reaches the first comparison threshold, a query is made as to whether to perform image result fusion. If the third comparison threshold is reached, it is considered to be the same target, for example, the same person, and a query is made as to whether to perform image result fusion.

[0051] In steps S2-S4 of the present embodiment, logical judgment of business process processing is mainly performed. In step S2, after configuration is completed, multi-image retrieval is performed. The multi-image retrieval in the present embodiment is different from the single-image retrieval based on the image content search method, that is, the retrieval of multiple images. When there are more than two images, pairwise calculations are performed. When multi-image retrieval is performed, taking two images as an example, when the retrieval request is initiated, the similarity of the images to be processed is obtained, the images are analyzed, and the images are compared pairwise to see if they are similar (reaching the set threshold). If the images are similar, it is asked whether to merge the results. If not, separate queries are performed on the images and they are displayed separately.

[0052] In step S3 of this embodiment, taking two pictures as an example, when a retrieval request is initiated, the pictures will be parsed, and the pictures will be compared to see if they are similar. When the similarity comparison result reaches the first comparison threshold, the initial images are determined to be similar, and the first fusion processing is performed. The first fusion processing in this embodiment includes: judging whether to perform the first data fusion; if not, the images to be processed are respectively subjected to image query and image download; if fused, the images to be processed are deduplicated before being downloaded. For example, if the pictures are similar, the user is asked whether to perform data fusion. If not, separate pictures are queried and displayed separately. If the user chooses fusion, multiple queries are performed, and the similarity is sorted according to the picture similarity. The optimal value after deduplication of the return values ​​of the two requests is taken. In this embodiment, the TopK value before the similarity after deduplication sorting is taken, and TopK is the maximum return value of the query.

[0053] In step S4 of this embodiment, when the similarity comparison result of the image to be processed does not reach the first comparison threshold, the initial image is determined to be dissimilar, an image query is performed on the image to be processed, and the number of repeated images is obtained according to the address information of the image data in the query result. If the data repetition reaches the second comparison threshold, it is determined that the processing objects are the same target, and the second fusion process is performed. In this embodiment, if the images are not similar, multiple queries are performed, and it is determined how many similar images there are in the query results.

[0054] In this embodiment, an image query is performed on the image to be processed; the address information of the image data in the query result is stored in a set with non-repeating elements, and the number of repeated images is obtained according to the length of the set; the data repetition degree is obtained according to the number of repeated images, and when the data repetition degree reaches the second comparison threshold, the second fusion processing is performed. The second fusion processing includes: determining whether to perform the second data fusion; if not, downloading the image to be processed; if fused, deduplicating the image to be processed before downloading the image. In this embodiment, the address information includes URL (Uniform Resource Locator), and multiple image queries are performed on the image to be processed; after obtaining the query result, the uniform resource locator of the image in each data is stored in a set with non-repeating elements; according to the length of the set and the total number of actual retrieval requests, the number of repeated images is obtained, and then the repetition degree is obtained. In this embodiment, each image is queried, and after obtaining the query result, the URL of the image in each data is taken out and stored in a set, and then the number of repeated images is determined according to the length of the set and the total number of actual two requests. The set in this embodiment is a set that supports efficient keyword query operations. It can check whether each given keyword is in the set, and supports efficient insertion and deletion. The Set does not guarantee that the data is in the same order as when it is stored when it is retrieved, and does not allow null values ​​or duplicate values. Since only non-duplicate data can be stored in the set, the total number of queries minus the length of the set can be used to obtain duplicate data. When downloading pictures, the pictures are rendered according to the values ​​in the set. In this way, it is possible to avoid the situation where a picture is downloaded multiple times.

[0055] In one embodiment, if two images are similar and the user chooses not to merge the data, the query status of each image is displayed separately. When the page loads the image data, the image is loaded according to the downloaded image data of the set. If the user chooses to merge the data, the data is sorted according to the similarity, and then the entire data is deduplicated, and the top K items are taken. If there are no K items, all the deduplicated data is returned, and the data is not displayed separately.

[0056] In one embodiment, if two pictures are not similar, pictures are first queried, and then the number of similar pictures is calculated. The calculation method in this embodiment is: the total number of the two queries minus the length of the set is repeated data, and the data duplication is calculated. The calculation method of the data duplication is: the repeated data is divided by the relatively long query data. For example, if there are 5 repeated items, the result of query A is 15, and the result of query B is 20, then the data duplication is:

[0057] 5 / 20*100%=25%,

[0058] If the similarity reaches the threshold, it is continued to determine whether data fusion is required.

[0059] The following is a specific example for explanation:

[0060] Example 1

[0061] like Figure 2 As shown, in this embodiment, a first comparison threshold, ie, 90% of the image feature similarity of the picture, a second comparison threshold, ie, 50% of the image data result similarity, is preset, and TopK is equal to 100.

[0062] First, upload two pictures and perform picture analysis. After analysis, use features to calculate similarity. If the image feature similarity of the pictures is higher than 90%, the user is asked whether to merge the results.

[0063] If the user does not choose to merge the results, two requests will be made, and the image address returned by the request will be placed in the set collection. The image will then be downloaded, and the page will be rendered after the download is complete.

[0064] If the image feature similarity of the image is higher than 90% and the user chooses data result fusion, the address of the image is first put into the set. If the length of the set is greater than TopK: 100, the top 100 image feature similarity data are selected for screening and the 100 image data are downloaded at the same time. If it is less than TopK: 100, all the image data after deduplication are directly downloaded for image download, and the page is rendered after the download is completed.

[0065] If the image feature similarity of the picture is lower than 90%, the user will not be asked whether to merge the results. Two requests for the picture will be made directly, and the results will be stored in the set. The length of the set will be determined (here, two requests for pictures are taken, and each request can query 100 results. The length of the set is 130). From the above calculation method, it can be seen that the repeated picture data is 100+100-130=70, and the repetition is: 70 / 100*100%=70%. If the data repetition is higher than 50%, the user will be asked whether to merge. If the data is merged, the data will be sorted, deduplicated, and the first 100 data will be filtered to download the pictures. Only the first 100 pictures after sorting and deduplication are downloaded. If the user does not merge the data, all pictures in the set will be downloaded, and the download will be 130 data.

[0066] Accordingly, this embodiment also provides a target judgment processing system based on multi-image retrieval data fusion, including:

[0067] A configuration module, used to pre-set a first comparison threshold for image feature similarity comparison and a second comparison threshold for data result similarity comparison;

[0068] Image acquisition module, used to collect image information,

[0069] A retrieval module is used to perform multi-image retrieval on the image to be processed according to the retrieval request and obtain the similarity of the image to be processed;

[0070] A processing module, configured to determine that the initial images are similar and perform a first fusion process when the similarity comparison result of the image to be processed reaches the first comparison threshold;

[0071] When the similarity comparison result of the image to be processed does not reach the first comparison threshold, it is determined that the initial image is not similar, an image query is performed on the image to be processed, and the number of repeated images is obtained based on the address information of the image data in the query result. If the data repetition reaches the second comparison threshold, it is determined that the processing objects are the same target, and a second fusion process is performed.

[0072] In this embodiment, first, a first comparison threshold for image feature similarity comparison and a second comparison threshold for data result similarity comparison are configured, as well as a third comparison threshold for indicating identical data. The third comparison threshold is greater than the first comparison threshold. If the similarity threshold reaches the first comparison threshold, a question is asked whether to perform image result fusion. If the third comparison threshold is reached, it is considered to be the same target, for example, the same person, and a question is asked whether to perform image result fusion.

[0073] In this embodiment, after the configuration is completed, a multi-image search is performed. The multi-image search in this embodiment is different from the single-image search based on the image content search method, that is, the search of multiple images. When there are more than two images, pairwise calculations are performed. Taking two images as an example, when a search request is initiated, the images will be parsed, and the images will be compared to see if they are similar. When the similarity comparison result reaches the first comparison threshold, the initial images are determined to be similar, and the first fusion processing is performed. The first fusion processing in this embodiment includes: determining whether to perform the first data fusion; if not, the images to be processed are respectively subjected to image query and image download; if fused, the images to be processed are deduplicated before being downloaded. For example, if the images are similar, the user is asked whether to perform data fusion. If not, separate images are queried and displayed separately. If the user chooses fusion, multiple queries are performed, and the similarity is sorted according to the image similarity. The optimal value obtained after deduplication of the return values ​​of the two requests is taken. In this embodiment, the TopK value before the similarity after deduplication sorting is taken, and TopK is the maximum return value of the query.

[0074] In this embodiment, when the similarity comparison result of the image to be processed does not reach the first comparison threshold, the initial image is determined to be dissimilar, an image query is performed on the image to be processed, and the number of repeated images is obtained according to the address information of the image data in the query result. If the data repetition reaches the second comparison threshold, it is determined that the processing objects are the same target, and the second fusion process is performed. In this embodiment, if the images are not similar, multiple queries are performed, and the number of similar images in the query results is determined.

[0075] In this embodiment, an image query is performed on the image to be processed; the address information of the image data in the query result is stored in a set with non-repeating elements, and the number of repeated images is obtained according to the length of the set; the data repetition degree is obtained according to the number of repeated images, and when the data repetition degree reaches the second comparison threshold, the second fusion processing is performed. The second fusion processing includes: determining whether to perform the second data fusion; if not, downloading the image to be processed; if fused, deduplicating the image to be processed before downloading the image. In this embodiment, the address information includes URL (Uniform Resource Locator), and multiple image queries are performed on the image to be processed; after obtaining the query result, the uniform resource locator of the image in each data is stored in a set with non-repeating elements; according to the length of the set and the total number of actual retrieval requests, the number of repeated images is obtained; when the image is downloaded, the image rendering is performed according to the data in the set. In this embodiment, each image is queried, and after obtaining the query result, the URL of the image in each data is taken out and stored in a set, and then the number of repeated images is determined according to the length of the set and the total number of actual two requests. The set in this embodiment is a set that supports efficient keyword query operations. It can check whether each given keyword is in the set, and supports efficient insertion and deletion. The Set does not guarantee that the data is in the same order as when it is stored when it is retrieved, and does not allow null values ​​or duplicate values. Since only non-duplicate data can be stored in the set, the total number of queries minus the length of the set can be used to obtain duplicate data. When downloading pictures, the pictures are rendered according to the values ​​in the set. In this way, it is possible to avoid the situation where a picture is downloaded multiple times.

[0076] In one embodiment, if two images are similar and the user chooses not to merge the data, the query status of each image is displayed separately. When the page loads the image data, the image is loaded according to the downloaded image data of the set. If the user chooses to merge the data, the data is sorted according to the similarity, and then the entire data is deduplicated, and the top K items are taken. If there are less than K items, all the deduplicated data is returned, and the data is not displayed separately.

[0077] The system in this embodiment, through the method in the above embodiment, can, on the one hand, determine the amount of duplicate data by the length of the set collection and the actual returned data, thereby greatly reducing the cost of deduplication. On the other hand, by determining whether data fusion is performed, if fusion is performed, the image addresses of the images with deduplication are sorted for download. If data fusion is not performed, all the image addresses of the elements in all the set collections fed back by the two query requests are used for image download, thereby saving protected bandwidth, memory and CPU resources.

[0078] This embodiment further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, any one of the methods in this embodiment is implemented.

[0079] This embodiment also provides an electronic terminal, including: a processor and a memory;

[0080] The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the terminal executes any one of the methods in this embodiment.

[0081] The computer-readable storage medium in this embodiment can be understood by ordinary technicians in this field: all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to the computer program. The aforementioned computer program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other media that can store program codes.

[0082] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication with each other. The memory is used to store computer programs, the communication interface is used to communicate, and the processor and the transceiver are used to run computer programs so that the electronic terminal executes each step of the above method.

[0083] In this embodiment, the memory may include a random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0084] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0085] In the above embodiments, unless otherwise specified, the use of serial numbers such as "first", "second", etc. to describe common objects only indicates that they refer to different instances of the same object, rather than indicating that the objects being described must be in a given order, whether temporally, spatially, sequentially, or in any other way. In the above embodiments, the description's reference to "this embodiment", "one embodiment", "another embodiment", or "other embodiments" indicates that the specific features, structures, or characteristics described in conjunction with the embodiments are included in at least some embodiments, but not necessarily all embodiments. Multiple occurrences of "this embodiment", "one embodiment", "another embodiment" do not necessarily all refer to the same embodiment.

[0086] In the above-described embodiments, although the invention has been described in conjunction with specific embodiments of the invention, many substitutions, modifications, and variations of these embodiments will be apparent to those of ordinary skill in the art based on the foregoing description. For example, other storage structures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed. Embodiments of the invention are intended to encompass all such substitutions, modifications, and variations that fall within the broad scope of the appended claims.

[0087] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0088] The present invention can be used in many general or special computing system environments or configurations, such as personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like.

[0089] The present invention may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0090] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present invention. Therefore, all equivalent modifications or changes made by a person of ordinary skill in the art without departing from the spirit and technical ideas disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A target judgment processing method based on multi-image retrieval data fusion, characterized in that: include: Presetting a first comparison threshold for image feature similarity comparison and a second comparison threshold for data result similarity comparison; Collect image information, perform multi-image retrieval on the image to be processed according to the retrieval request, and obtain the similarity of the image to be processed; When the similarity comparison result of the image to be processed reaches the first comparison threshold, the initial images are determined to be similar, and a first fusion process is performed; The first fusion processing includes: determining whether to perform first data fusion; if not, respectively performing image query and image download on the image to be processed; if fusion, performing deduplication processing on the image to be processed before downloading the image; When the similarity comparison result of the image to be processed does not reach the first comparison threshold, it is determined that the initial image is not similar, an image query is performed on the image to be processed, and the number of repeated images is obtained according to the address information of the image data in the query result. If the data repetition reaches the second comparison threshold, it is determined that the processing objects are the same target, and a second fusion process is performed; Performing an image query on the image to be processed, and obtaining the number of repeated images according to the address information of the image data in the query result includes: performing an image query on the image to be processed; storing the address information of the image data in the query result into a set whose elements are not repeated, and obtaining the number of repeated images according to the length of the set; The second fusion processing includes: determining whether to perform second data fusion; if not, downloading the image to be processed; if fusion is performed, deduplication processing is performed on the image to be processed before downloading the image.

2. The target judgment processing method based on multi-image retrieval data fusion according to claim 1 is characterized in that: When the similarity comparison result of the image to be processed does not reach the first comparison threshold, determining that the initial image is not similar includes: The data repetition degree is acquired according to the number of repeated images, and the second fusion process is performed when the data repetition degree reaches a second comparison threshold.

3. The target judgment processing method based on multi-image retrieval data fusion according to claim 2 is characterized in that: The address information includes a uniform resource locator, Performing multiple image queries on the image to be processed; After obtaining the query results, store the uniform resource locator of the image in each piece of data into a set with no duplicate elements; According to the length of the set and the total number of actual search requests, the number of repeated images is obtained, and then the data duplication degree is obtained.

4. The target judgment processing method based on multi-image retrieval data fusion according to claim 3 is characterized in that: The first fusion process further includes: If not, the query results of each image are displayed separately, and when loading image data, the image is loaded according to the downloaded image data of the set; If fusion is performed, multiple image queries are performed on the image to be processed, and the images are sorted according to their similarity. After all data are deduplicated, the optimal value is taken for feedback.

5. The target judgment processing method based on multi-image retrieval data fusion according to claim 4 is characterized in that: Acquire duplicate data according to the difference between the total number of data in the image query and the length of the set; When fusion is selected in the first fusion process, the pictures are sorted according to the similarity of image features. After all data are deduplicated, the top K data are fed back. If there are less than K data, all the deduplicated data are fed back. The data duplication in the second fusion process is obtained by the ratio of the repeated data to the query data.

6. The target judgment processing method based on multi-image retrieval data fusion according to claim 1 is characterized in that: Presetting a third comparison threshold for indicating identical data, wherein the third comparison threshold is greater than the first comparison threshold; When the similarity comparison result of the images to be processed reaches a third comparison threshold, it is determined that the processing objects are the same target, and the second fusion processing is performed.

7. A target judgment processing system based on multi-image retrieval data fusion, characterized in that: include: A configuration module, used to pre-set a first comparison threshold for image feature similarity comparison and a second comparison threshold for data result similarity comparison; Image acquisition module, used to collect image information, A retrieval module is used to perform multi-image retrieval on the image to be processed according to the retrieval request and obtain the similarity of the image to be processed; A processing module, configured to determine that the initial images are similar and perform a first fusion process when the similarity comparison result of the image to be processed reaches the first comparison threshold; When the similarity comparison result of the image to be processed does not reach the first comparison threshold, it is determined that the initial image is not similar, an image query is performed on the image to be processed, and the number of repeated images is obtained according to the address information of the image data in the query result. If the data repetition reaches the second comparison threshold, it is determined that the processing objects are the same target, and a second fusion process is performed; The first fusion processing includes: determining whether to perform first data fusion; if not, respectively performing image query and image download on the image to be processed; if fusion, performing deduplication processing on the image to be processed before downloading the image; Performing an image query on the image to be processed, and obtaining the number of repeated images according to the address information of the image data in the query result includes: performing an image query on the image to be processed; storing the address information of the image data in the query result into a set whose elements are not repeated, and obtaining the number of repeated images according to the length of the set; The second fusion processing includes: determining whether to perform second data fusion; if not, downloading the image to be processed; if fusion is performed, deduplication processing is performed on the image to be processed before downloading the image.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

9. An electronic terminal, characterized in that: include: Processor and memory; The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the terminal executes the method according to any one of claims 1 to 6.