An image deduplication method, device, equipment and readable storage medium

Abstract

The present application relates to the field of image processing, and more particularly to an image deduplication method, device, equipment and readable storage medium, the method acquires first information, the first information includes at least two frames of image information obtained by camera;The first information is sent to the preprocessing module for coordinate conversion and target identification frame selection processing, and second information is obtained, the second information includes the endpoint coordinate value information of all target frames identified in the first information;The second information is sent to the matching module for matching;The endpoint coordinate value of the corresponding target frame in the first image and the second image is subjected to IOU screening, center offset screening and score value screening, and target image information in a stationary state is obtained;Based on the image information of the target state in a stationary state, the target image in the first information is deduplicated to obtain deduplicated image information.The present application removes the stationary target to achieve the purpose of target deduplication.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and in particular to an image deduplication method, apparatus, device and readable storage medium. Background Art

[0002] As the AI ​​algorithm industry continues to evolve, it's becoming increasingly common for managers to find that, in practical AI applications, they no longer need to continuously monitor the same static or slightly moving target under the same camera, as this wastes supervisory resources. To address this, a method is needed to remove static or slightly moving targets. Summary of the Invention

[0003] The present invention aims to provide an image deduplication method, apparatus, device, and readable storage medium to improve the above-mentioned problem. To achieve the above-mentioned object, the present invention adopts the following technical solutions:

[0004] On the one hand, the present application provides an image deduplication method, which includes: obtaining first information, the first information including at least two frames of image information obtained by a camera; sending the first information to a preprocessing module for coordinate conversion and target recognition and frame selection processing to obtain second information, the second information including endpoint coordinate value information of all target frames identified in the first information; sending the second information to a matching module for matching to obtain the endpoint coordinate values ​​of the first image and the corresponding target frames in the second image in the second information, the first image being the previous frame image of the two adjacent frames in the second information, and the second image being the next frame image of the two adjacent frames in the second information; performing IOU filtering, center offset filtering and score value filtering on the endpoint coordinate values ​​of the corresponding target frames in the first image and the second image to obtain target image information in a stationary state; deduplicating the target image in the first information based on the image information in a stationary state to obtain deduplicated image information.

[0005] In a second aspect, an embodiment of the present application provides an image deduplication device, comprising:

[0006] A first acquiring unit, configured to acquire first information, where the first information includes at least two frames of image information acquired by a camera;

[0007] a first processing unit, configured to send the first information to a preprocessing module for coordinate conversion and target identification and selection processing to obtain second information, where the second information includes coordinate value information of endpoints of all target boxes identified in the first information;

[0008] The first matching unit is configured to send the second information to a matching module for matching to obtain endpoint coordinate values of corresponding target boxes in a first image and a second image in the second information, the first image being a previous frame image of two adjacent frames of images in the second information, and the second image being a next frame image of the two adjacent frames of images in the second information.

[0009] The second processing unit is configured to perform IOU screening, center offset screening and score value screening on the endpoint coordinate values of the corresponding target boxes in the first image and the second image to obtain target image information in a state of being static.

[0010] The third processing unit is configured to perform deduplication on target images in the first information based on the image information in the state of being static to obtain deduplicated image information.

[0011] In a third aspect, an embodiment of the present application provides an image deduplication device, the device comprising a memory and a processor. The memory is configured to store a computer program; and the processor is configured to execute the computer program to implement the steps of the image deduplication method.

[0012] In a fourth aspect, an embodiment of the present application provides a readable storage medium, the readable storage medium storing a computer program, the computer program being executed by a processor to implement the steps of the image deduplication method.

[0013] The present application has the following beneficial effects:

[0014] The present application matches the targets recognized by the algorithm models of the previous and next frames, and calculates the IOU and the center point offset of the matched targets. In order to avoid the influence of the algorithm model recognition error, a reasonable threshold of the IOU and the center point offset is designed to determine whether the current matched target is static, and the static target is removed to achieve the purpose of target deduplication.

[0015] Other features and advantages of the present application will be described in the following description, and some will become apparent from the description, or will be understood from the practice of the present application. The purpose and other advantages of the present application can be achieved and obtained by the structures specifically pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be considered as limiting the scope. For those skilled in the art, other related drawings can also be obtained from these drawings without creative labor.

[0017] Figure 1 Schematic diagram of a flow chart of an image deduplication method according to an embodiment of the present invention;

[0018] Figure 2 Schematic diagram of the structure of an image deduplication device according to an embodiment of the present invention;

[0019] Figure 3 This is a schematic diagram of the structure of an image deduplication device described in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0021] It should be noted that similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are used only to distinguish the description and should not be understood as indicating or implying relative importance.

[0022] like Figure 1 As shown, this embodiment provides an image deduplication method, which includes step S1, step S2, step S3, step S4 and step S5.

[0023] Step S1: Acquire first information, where the first information includes at least two frames of image information acquired by a camera;

[0024] Step S2: Send the first information to a preprocessing module for coordinate conversion and target identification and selection processing to obtain second information, where the second information includes coordinate value information of endpoints of all target boxes identified in the first information;

[0025] Step S3: Send the second information to a matching module for matching, and obtain the endpoint coordinate values ​​of the corresponding target frames in the first image and the second image in the second information, where the first image is the previous frame image of the two adjacent frames in the second information, and the second image is the next frame image of the two adjacent frames in the second information;

[0026] Step S4, the end point coordinate values of the corresponding target frame in the first image and the second image are subjected to IOU screening, center offset screening and score value screening, to obtain target image information in a stationary state;

[0027] Step S5, based on the image information of the target state in a stationary state, the target image in the first information is de-duplicated to obtain de-duplicated image information.

[0028] The application matches the targets recognized by the algorithm model in the front and rear two frames, and calculates the IOU and the center point offset of the matched target. In order to avoid the influence of the algorithm model recognition error, a reasonable threshold of IOU and center point offset is designed to determine whether the current matched target is stationary, and the stationary target is removed to achieve the purpose of target de-duplication.

[0029] In one specific embodiment of the present disclosure, the step S2 includes steps S21 and S22.

[0030] Step S21, the image information in the first information is sent to a two-dimensional rectangular space coordinate system for processing to obtain first sub-information, and the first sub-information is image information with determined coordinate values of each pixel point;

[0031] Step S22, based on a target detection algorithm, the first sub-information is subjected to target recognition processing, and all targets are framed to obtain the target frame of each target and the end point coordinate value information of all target frames.

[0032] The application processes the target to be recognized through coordinate conversion and target framing, and marks the end point coordinates of the target, thereby preparing for the subsequent association matching.

[0033] In one specific embodiment of the present disclosure, the step S3 includes steps S31, S32, S33 and S34.

[0034] Step S31, based on kalman filtering, all target frame end point coordinate value information in adjacent two frames of image information is processed, the target frame end point coordinates in the first image are traversed, the end point coordinate values of all target frame end point coordinate values in the first image in the second image are estimated by kalman filtering, to obtain the end point coordinate values of the estimated frame of all target frame coordinate values in the first image in the second image;

[0035] Step S32, based on the target frame end point coordinate values in the second image, the target frame center point coordinate values in the second image are obtained, and whether the target frame center point coordinate values in the second image are in the estimated frame is judged to obtain a first judgment result;

[0036] Step S33, if the first judgment result is that the target box center point coordinate value in the second image is in the estimated box, the IOU value of the end point coordinate value of the target box in all the second images and the end point coordinate value of the estimated box is calculated to obtain at least one first IOU value;

[0037] Step S34, based on at least one first IOU value, the end point coordinate value of the corresponding target box in the first image and the second image is obtained.

[0038] It can be understood that the above steps reduce the search range to a certain small range by introducing kalman filtering, and quickly determine the best matching target box.

[0039] In one specific embodiment of the present disclosure, the step S4 comprises steps S41, S42, S43 and S44.

[0040] Step S41, based on the end point coordinate value of the corresponding target box in the first image and the second image, a second IOU value is obtained, the second IOU value is the IOU value of the end point coordinate value of the first image target box and the end point coordinate value of the corresponding target box in the second image,

[0041] Step S42, the NMS error value in the target detection model is determined based on the error calculation formula, and the third IOU value is obtained by adding the NMS error value and the second IOU value;

[0042] Step S43, based on the comparison between the third IOU value and the preset IOU threshold value, if the third IOU value is less than the preset IOU threshold value, the target box coordinate corresponding to the third IOU value is deleted, and it is marked that the target in the target box corresponding to the third IOU value is in motion state;

[0043] Step S44, if the third IOU value is greater than or equal to the preset IOU threshold value, the target corresponding to the third IOU value is subjected to center offset screening and score value screening to obtain the target image information in static state.

[0044] It can be understood that the above steps calculate the IOU value of the two target boxes matched by using the matched target box. Since the threshold value of NMS affects the detection result when the target detection model NMS operates, the IOU value calculated in the last step is not accurate, so an error value is introduced to enhance the detection accuracy.

[0045] In one specific embodiment of the present disclosure, the step S44 comprises steps S441, S442 and S443.

[0046] Step S441: Mark a third IOU value that is greater than or equal to a preset IOU threshold, and process the target frame in two adjacent image frames corresponding to the marked third IOU value to obtain the center point coordinates of the target frame in the two adjacent image frames;

[0047] Step S442: Calculating the offset of the center point coordinates of the target frame in the two adjacent image frames to obtain the X-axis offset of the center point coordinates of the target frame in the two adjacent image frames and the Y-axis offset of the center point coordinates of the target frame in the two adjacent image frames;

[0048] Step S443: Compare the X-axis offset of the center point coordinates of the target frame in two adjacent frames and the Y-axis offset of the center point coordinates of the target frame in two adjacent frames with a preset offset threshold to obtain a stationary target image.

[0049] It can be understood that if the X-axis offset of the center point coordinate and the Y-axis offset of the center point coordinate are both smaller than the offset threshold, it can be determined that the image is a stationary target image.

[0050] It can be understood that the present application filters the third IOU value greater than the preset IOU threshold, and then calculates the center point of the target frame corresponding to the third IOU value greater than the preset IOU threshold, obtains the center point coordinates of the target frame in two adjacent frames, and calculates the position change of the center point of the target frame in the two adjacent frames based on the center point coordinates, including X-axis change and Y-axis change, and the offset threshold preset in the present invention can be adjusted based on the aspect ratio of the target frame and the overall image. If the aspect ratio is greater than one-half, the offset threshold can be adjusted upward, and if the aspect ratio is less than one-half, the offset threshold can be adjusted downward.

[0051] It can be understood that the calculation formula of the offset in the above steps is:

[0052]

[0053]

[0054] Where x represents the offset of the center point of the corresponding target frame in the two adjacent frames on the X axis, x1 represents the X axis coordinate of the center point of the target frame in the previous frame of the two adjacent frames, and x2 represents the X axis coordinate of the center point of the target frame in the next frame of the two adjacent frames; y represents the offset of the center point of the corresponding target frame in the two adjacent frames on the Y axis, y1 represents the Y axis coordinate of the center point of the target frame in the previous frame of the two adjacent frames, and y2 represents the Y axis coordinate of the center point of the target frame in the next frame of the two adjacent frames.

[0055] In a specific embodiment of the present disclosure, step S44 includes step S444, step S445 and step S446.

[0056] Step S444: Mark the third IOU value that is greater than or equal to the preset IOU threshold, and perform category analysis on the target in the target box corresponding to the marked third IOU value to obtain the target category in the target box corresponding to the third IOU value;

[0057] Step S445: determining whether the target category in the target box corresponding to the third IOU value changes; if the target category in the target box corresponding to the third IOU value does not change, determining that the target in the target box corresponding to the third IOU value is a stationary target;

[0058] It can be understood that determining the target category in the target frame corresponding to the third IOU value is to compare the target categories in the corresponding target frames of two adjacent frames, and then determine whether the category has changed.

[0059] Step S446: If the target category in the target box corresponding to the third IOU value does not change, it is determined that the target in the target box corresponding to the third IOU value is a moving target.

[0060] It can be understood that the above steps are to determine whether the targets in the same target frame are of the same category, so as to prevent two objects of the same size from moving, but the captured image obtains the target frame of the same size, and then the objects are mistakenly judged as not moving.

[0061] Example 2

[0062] like Figure 2 As shown, this embodiment provides an image deduplication device, which includes a first acquisition unit 701, a first processing unit 702, a first matching unit 703, a second processing unit 704 and a third processing unit 705.

[0063] A first acquiring unit 701 is configured to acquire first information, where the first information includes at least two frames of image information acquired by a camera;

[0064] A first processing unit 702 is configured to send the first information to a pre-processing module for coordinate conversion and target identification and selection processing to obtain second information, where the second information includes coordinate value information of endpoints of all target boxes identified in the first information;

[0065] A first matching unit 703 is configured to send the second information to a matching module for matching, and obtain endpoint coordinate values ​​of corresponding target frames in a first image and a second image in the second information, where the first image is a previous frame image between two adjacent frames in the second information, and the second image is a subsequent frame image between two adjacent frames in the second information;

[0066] The second processing unit 704 is configured to perform IOU filtering, center offset filtering, and score filtering on the endpoint coordinates of the corresponding target boxes in the first image and the second image to obtain stationary target image information;

[0067] The third processing unit 705 is configured to deduplicate the target image in the first information based on the image information in which the target state is still, to obtain deduplicated image information.

[0068] In a specific embodiment of the present disclosure, the first processing unit 702 includes a first processing sub-unit 7021 and a second processing sub-unit 7022 .

[0069] The first processing sub-unit 7021 is configured to send the image information in the first information to a two-dimensional rectangular space coordinate system for processing to obtain first sub-information, where the first sub-information is image information with the coordinate value of each pixel point determined;

[0070] The second processing sub-unit 7022 is used to perform target recognition processing on the first sub-information based on the target detection algorithm, select all targets, and obtain the target frame of each target and the endpoint coordinate value information of all target frames.

[0071] In a specific embodiment of the present disclosure, the first matching unit 703 includes a third processing subunit 7031 , a first judgment subunit 7032 , a second judgment subunit 7033 and a fourth processing subunit 7034 .

[0072] The third processing sub-unit 7031 is used to process the target frame endpoint coordinate value information in two adjacent frames of image information based on Kalman filtering, traverse the target frame endpoint coordinates in the first image, estimate the endpoint coordinate values ​​of all target frame endpoint coordinate values ​​in the first image in the second image through Kalman filtering, and obtain the endpoint coordinate values ​​of the estimated frame of all target frame coordinate values ​​in the first image in the second image;

[0073] The first judgment subunit 7032 is configured to obtain the coordinate values ​​of the center point of the target frame in the second image based on the coordinate values ​​of the endpoints of the target frame in the second image, and to determine whether the coordinate value of the center point of the target frame in the second image is within the estimation frame, thereby obtaining a first judgment result.

[0074] The second judgment subunit 7033 is configured to, if the first judgment result is that the coordinate value of the center point of the target frame in the second image is within the estimated frame, calculate the IOU values ​​of the endpoint coordinate values ​​of all target frames in the second image and the endpoint coordinate values ​​of the estimated frame to obtain at least one first IOU value;

[0075] The fourth processing sub-unit 7034 is configured to obtain endpoint coordinate values ​​of corresponding target boxes in the first image and the second image based on at least one of the first IOU values.

[0076] In a specific embodiment of the present disclosure, the second processing unit 704 includes a first calculation subunit 7041 , a second calculation subunit 7042 , a third judgment subunit 7043 and a fourth judgment subunit 7044 .

[0077] The first calculation subunit 7041 is configured to obtain a second IOU value based on the endpoint coordinate values ​​of the corresponding target boxes in the first image and the second image, where the second IOU value is the IOU value of the endpoint coordinate values ​​of the target box in the first image and the endpoint coordinate values ​​of the corresponding target box in the second image.

[0078] A second calculation subunit 7042 is configured to determine an NMS error value within the object detection model based on an error calculation formula, and add the NMS error value to the second IOU value to obtain a third IOU value;

[0079] The third judgment subunit 7043 is configured to compare the third IOU value with a preset IOU threshold. If the third IOU value is less than the preset IOU threshold, the target frame coordinates corresponding to the third IOU value are deleted, and the target in the target frame corresponding to the third IOU value is marked as being in motion.

[0080] The fourth judgment subunit 7044 is configured to perform center offset screening and score value screening on the target corresponding to the third IOU value if the third IOU value is greater than or equal to a preset IOU threshold, so as to obtain image information of the target in a stationary state.

[0081] In a specific embodiment of the present disclosure, the fourth judgment subunit 7044 includes a first marking subunit 70441 , a third calculation subunit 70442 and a fifth processing subunit 70443 .

[0082] The first marking subunit 70441 is used to mark a third IOU value that is greater than or equal to a preset IOU threshold, and process the target frame in two adjacent frames corresponding to the marked third IOU value to obtain the center point coordinates of the target frame in the two adjacent frames;

[0083] The third calculation subunit 70442 is used to calculate the offset of the center point coordinates of the target frame in the two adjacent frames of image to obtain the X-axis offset of the center point coordinates of the target frame in the two adjacent frames of image and the Y-axis offset of the center point coordinates of the target frame in the two adjacent frames of image;

[0084] The fifth processing sub-unit 70443 is used to compare the X-axis offset of the center point coordinates of the target frame in two adjacent frame images and the Y-axis offset of the center point coordinates of the target frame in two adjacent frame images with a preset offset threshold to obtain a stationary target image.

[0085] In a specific embodiment of the present disclosure, the fourth judgment subunit 7044 includes a second marking subunit 70444 , a fifth judgment subunit 70445 and a sixth judgment subunit 70446 .

[0086] The second marking subunit 70444 is used to mark the third IOU value that is greater than or equal to the preset IOU threshold, and perform category analysis on the target in the target box corresponding to the marked third IOU value to obtain the target category in the target box corresponding to the third IOU value;

[0087] The fifth judgment subunit 70445 is used to judge whether the target category in the target box corresponding to the third IOU value changes; if the target category in the target box corresponding to the third IOU value does not change, then the target in the target box corresponding to the third IOU value is judged to be a stationary target;

[0088] The sixth judgment subunit 70446 is configured to determine that the target in the target box corresponding to the third IOU value is a moving target if the target category in the target box corresponding to the third IOU value does not change.

[0089] It should be noted that, regarding the apparatus in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated on here.

[0090] Example 3

[0091] Corresponding to the above method embodiment, the embodiment of the present disclosure further provides an image deduplication device. The image deduplication device described below and the image deduplication method described above can refer to each other.

[0092] Figure 3 FIG. 8 is a block diagram of an image deduplication device 800 according to an exemplary embodiment. Figure 3As shown, the image deduplication device 800 may include: a processor 801 and a memory 802. The image deduplication device 800 may also include one or more of a multimedia component 803, an input / output (I / O) interface 804, and a communication component 805.

[0093] The processor 801 is used to control the overall operation of the image deduplication device 800 to complete all or part of the steps in the above-mentioned image deduplication method. The memory 802 is used to store various types of data to support the operation of the image deduplication device 800. Such data may include, for example, instructions for any application or method operating on the image deduplication device 800, as well as application-related data, such as contact information, sent and received messages, pictures, audio, video, etc. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may be further stored in the memory 802 or sent through the communication component 805. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 804 provides an interface between the processor 801 and other interface modules. The above-mentioned other interface modules may be a keyboard, a mouse, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 805 is used for wired or wireless communication between the image deduplication device 800 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, near field communication (NFC), 2G, 3G or 4G, or a combination of one or more thereof, so the corresponding communication component 805 may include: a Wi-Fi module, a Bluetooth module, an NFC module.

[0094] In an exemplary embodiment, the image deduplication device 800 can be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-mentioned image deduplication method.

[0095] In another exemplary embodiment, a computer-readable storage medium including program instructions is further provided. When executed by a processor, the program instructions implement the steps of the above-described image deduplication method. For example, the computer-readable storage medium may be the aforementioned memory 802 including the program instructions. The program instructions may be executed by the processor 801 of the image deduplication device 800 to implement the above-described image deduplication method.

[0096] Example 4

[0097] Corresponding to the above method embodiment, the embodiment of the present disclosure further provides a readable storage medium. The readable storage medium described below and the image deduplication method described above can refer to each other.

[0098] A readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the image deduplication method of the above method embodiment.

[0099] The readable storage medium may specifically be any readable storage medium that can store program code, 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.

[0100] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

[0101] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for deduplication of an image, characterized in that: include: Acquire first information, where the first information includes at least two frames of image information acquired by a camera; Sending the first information to a preprocessing module for coordinate conversion and target identification and selection processing to obtain second information, where the second information includes coordinate value information of endpoints of all target boxes identified in the first information; Sending the second information to a matching module for matching, obtaining endpoint coordinate values ​​of target frames corresponding to the first image and the second image in the second information, where the first image is the previous frame of the two adjacent frames in the second information, and the second image is the next frame of the two adjacent frames in the second information; Performing IOU screening, center offset screening, and score screening on the endpoint coordinates of the corresponding target boxes in the first image and the second image to obtain stationary target image information; Deduplicating the target image in the first information based on the image information in which the target state is still, to obtain deduplicated image information; The second information is sent to the matching module for matching, and the endpoint coordinate values ​​of the corresponding target frames in the first image and the second image in the second information are obtained, including: Based on the Kalman filter, all target frame endpoint coordinate value information in two adjacent frames of image information is processed, the target frame endpoint coordinates in the first image are traversed, and the endpoint coordinate values ​​of all target frame endpoint coordinate values ​​in the first image in the second image are estimated by the Kalman filter to obtain the endpoint coordinate values ​​of the estimated frame of all target frame coordinate values ​​in the first image in the second image; Obtaining a center point coordinate value of the target frame in the second image based on the endpoint coordinate values ​​of the target frame in the second image, and determining whether the center point coordinate value of the target frame in the second image is within the estimation frame to obtain a first determination result; If the first judgment result is that the coordinate value of the center point of the target frame in the second image is within the estimated frame, calculating the IOU values ​​of the endpoint coordinate values ​​of all target frames in the second image and the endpoint coordinate values ​​of the estimated frame to obtain at least one first IOU value; Based on at least one of the first IOU values, obtaining endpoint coordinate values ​​of corresponding target boxes in the first image and the second image; The endpoint coordinate values ​​of the corresponding target boxes in the first image and the second image are subjected to IOU screening, center offset screening, and score value screening, including: Based on the endpoint coordinate values ​​of the corresponding target boxes in the first image and the second image, a second IOU value is obtained, where the second IOU value is the IOU value of the endpoint coordinate values ​​of the target box in the first image and the endpoint coordinate values ​​of the corresponding target box in the second image. Determine an NMS error value within the target detection algorithm based on an error calculation formula, and add the NMS error value to the second IOU value to obtain a third IOU value; Based on the comparison between the third IOU value and a preset IOU threshold, if the third IOU value is less than the preset IOU threshold, the target frame coordinates corresponding to the third IOU value are deleted, and the target in the target frame corresponding to the third IOU value is marked as being in motion; If the third IOU value is greater than or equal to a preset IOU threshold, center offset screening and score value screening are performed on the target corresponding to the third IOU value to obtain image information of the stationary target.

2. The image deduplication method according to claim 1, wherein: The first information is sent to a preprocessing module for coordinate conversion and target recognition and selection processing to obtain second information, including: Sending the image information in the first information to a two-dimensional rectangular space coordinate system for processing to obtain first sub-information, where the first sub-information is image information with the coordinate value of each pixel point determined; The first sub-information is subjected to target recognition processing based on a target detection algorithm, all targets are framed, and a target frame of each target and endpoint coordinate value information of all target frames are obtained.

3. An image deduplication device, characterized in that: include: A first acquiring unit, configured to acquire first information, where the first information includes at least two frames of image information acquired by a camera; a first processing unit, configured to send the first information to a preprocessing module for coordinate conversion and target identification and selection processing to obtain second information, where the second information includes coordinate value information of endpoints of all target boxes identified in the first information; a first matching unit, configured to send the second information to a matching module for matching, to obtain endpoint coordinate values ​​of a target frame corresponding to a first image and a second image in the second information, where the first image is a previous frame image of two adjacent frames in the second information, and the second image is a subsequent frame image of two adjacent frames in the second information; a second processing unit, configured to perform IOU filtering, center offset filtering, and score value filtering on the endpoint coordinate values ​​of the corresponding target boxes in the first image and the second image to obtain stationary target image information; a third processing unit, configured to deduplicate the target image in the first information based on the image information in which the target state is still, to obtain deduplicated image information; Wherein, the first matching unit includes: The third processing subunit is configured to process all target frame endpoint coordinate value information in two adjacent frames of image information based on a Kalman filter, traverse the target frame endpoint coordinates in the first image, estimate the endpoint coordinate values ​​of all target frame endpoint coordinate values ​​in the first image in the second image through a Kalman filter, and obtain the endpoint coordinate values ​​of the estimated frame of all target frame coordinate values ​​in the first image in the second image; a first judgment subunit, configured to obtain a coordinate value of a center point of a target frame in the second image based on the coordinate values ​​of endpoints of the target frame in the second image, and to determine whether the coordinate value of the center point of the target frame in the second image is within an estimation frame, thereby obtaining a first judgment result; a second judgment subunit, configured to, if the first judgment result is that the coordinate value of the center point of the target frame in the second image is within the estimated frame, calculate the IOU values ​​of the endpoint coordinate values ​​of all target frames in the second image and the endpoint coordinate values ​​of the estimated frame to obtain at least one first IOU value; a fourth processing subunit, configured to obtain endpoint coordinate values ​​of corresponding target boxes in the first image and the second image based on at least one of the first IOU values; Wherein, the second processing unit includes: The first calculation subunit is configured to obtain a second IOU value based on the endpoint coordinate values ​​of the corresponding target boxes in the first image and the second image, where the second IOU value is the IOU value of the endpoint coordinate values ​​of the target box in the first image and the endpoint coordinate values ​​of the corresponding target box in the second image. A second calculation subunit is configured to determine an NMS error value within the target detection algorithm based on an error calculation formula, and add the NMS error value to the second IOU value to obtain a third IOU value; a third judgment subunit, configured to compare the third IOU value with a preset IOU threshold, and if the third IOU value is less than the preset IOU threshold, delete the target frame coordinates corresponding to the third IOU value, and mark the target in the target frame corresponding to the third IOU value as being in motion; The fourth judgment subunit is configured to perform center offset screening and score value screening on the target corresponding to the third IOU value if the third IOU value is greater than or equal to a preset IOU threshold, so as to obtain image information of the target in a stationary state.

4. The image deduplication device according to claim 3, characterized in that: The device comprises: a first processing sub-unit, configured to send the image information in the first information to a two-dimensional rectangular space coordinate system for processing to obtain first sub-information, where the first sub-information is image information with the coordinate value of each pixel point determined; The second processing sub-unit is used to perform target recognition processing on the first sub-information based on the target detection algorithm, select all targets, and obtain the target frame of each target and the endpoint coordinate value information of all target frames.

5. An image deduplication device, characterized in that: include: memory for storing computer programs; A processor, configured to implement the steps of the image deduplication method according to any one of claims 1 to 2 when executing the computer program.

6. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the image deduplication method according to any one of claims 1 to 2 are implemented.

Images