Target person trajectory confirmation method and device, equipment and medium
By performing local small-area feature similarity analysis and feature type screening on video points in pedestrian re-identification, the accuracy of pedestrian trajectory confirmation is solved, and the accuracy of trajectory confirmation is improved.
Patent Information
- Application Number
- CN202311407529.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-27
- Publication Date
- 2025-04-29
AI Technical Summary
In the prior art, it is difficult to accurately determine the true trajectory of pedestrians in pedestrian re-identification, especially when similar human bodies appear at multiple points.
Through the local small area features and their feature types of target personnel, the suspected points in the video are checked, the trajectory points and candidate suspected points are determined, the similarity of the local small area feature is calculated, and the trajectory of the target personnel is confirmed based on the feature type screening results.
It effectively solves the problem of trajectory points confusion in pedestrian re-identification and improves the accuracy of target personnel's trajectory confirmation.
Smart Images
Figure CN119904904A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of target detection technology, and in particular to a method, device, equipment and medium for confirming a trajectory of a target person. Background Art
[0002] With the widespread use of smart cameras, pedestrian re-identification has been increasingly used. However, the accuracy of pedestrian re-identification has always been a major problem, resulting in deviations in the analysis of pedestrian trajectories. When similar human bodies appear at multiple locations and their temporal and spatial positions meet the requirements, it is impossible to accurately determine the pedestrian's true trajectory.
[0003] In a related scheme, a secondary search comparison is performed directly in the search results through local features to further confirm the human target. However, in this scheme, not all search targets are obviously different, so accurate search cannot be performed. Summary of the invention
[0004] The present invention provides a method, device, equipment and medium for confirming the trajectory of a target person. Based on the local small area features of the target person and its feature type, suspected points in the video are checked, which can effectively solve the problem of trajectory point confusion in pedestrian re-identification, thereby improving the accuracy of target person trajectory confirmation.
[0005] According to one aspect of the present invention, a method for confirming a target person's trajectory is provided, the method comprising:
[0006] Determine the confirmed trajectory point and at least two candidate suspected points from the video to be checked based on the original feature information of the target person;
[0007] Determine the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, and determine the screening result of the candidate suspected point according to the local small area feature similarity and the feature type of the local small area feature;
[0008] The trajectory confirmation result of the target person is determined according to the screening result of the candidate suspected points.
[0009] According to another aspect of the present invention, a target person's trajectory confirmation device is provided, comprising:
[0010] A point determination module is used to determine the confirmed trajectory point and at least two candidate suspected points from the video to be checked according to the original feature information of the target person;
[0011] A candidate suspected point screening module is used to respectively determine the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, and determine the screening result of the candidate suspected point according to the local small area feature similarity and the feature type of the local small area feature;
[0012] The trajectory confirmation result determination module is used to determine the trajectory confirmation result of the target person according to the screening result of the candidate suspected points.
[0013] According to another aspect of the present invention, there is provided an electronic device, the electronic device comprising:
[0014] at least one processor; and
[0015] a memory communicatively connected to the at least one processor; wherein,
[0016] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the target person's trajectory confirmation method described in any embodiment of the present invention.
[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the target person's trajectory confirmation method described in any embodiment of the present invention when executed.
[0018] The technical solution of the embodiment of the present invention determines the confirmed trajectory point and at least two candidate suspected points from the video to be checked based on the original feature information of the target person; respectively determines the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, and determines the screening result of the candidate suspected points based on the local small area feature similarity and the feature type of the local small area feature; determines the trajectory confirmation result of the target person based on the screening result of the candidate suspected points. This technical solution, which checks the suspected points in the video based on the local small area features of the target person and their feature types, can effectively solve the problem of trajectory point confusion in pedestrian re-identification, thereby improving the accuracy of target person trajectory confirmation.
[0019] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0021] Figure 1 is a flow chart of a method for confirming a target person's trajectory provided according to Embodiment 1 of the present invention;
[0022] Figure 2 This is a schematic diagram of a target person's trajectory provided according to Embodiment 1 of the present invention;
[0023] Figure 3 is a flow chart of a method for confirming a target person's trajectory provided according to Embodiment 2 of the present invention;
[0024] Figure 4 is a schematic diagram of target local feature segmentation provided according to Embodiment 2 of the present invention;
[0025] Figure 5 is a flow chart of a method for confirming a target person's trajectory provided according to Embodiment 3 of the present invention;
[0026] Figure 6 is a schematic diagram of determining a person's direction according to Embodiment 3 of the present invention;
[0027] Figure 7 2 is a schematic diagram of the structure of a target person's trajectory confirmation device provided according to a fourth embodiment of the present invention;
[0028] Figure 8 The present invention is a schematic diagram of the structure of an electronic device for implementing a method for confirming a target person's trajectory according to an embodiment of the present invention. DETAILED DESCRIPTION
[0029] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme 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 only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0030] It should be noted that the terms "first", "second", "target", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0031] Embodiment 1
[0032] Figure 1 This is a flow chart of a method for confirming the trajectory of a target person provided in the first embodiment of the present invention. This embodiment is applicable to the case of quickly and accurately confirming the trajectory of a target person in a video. The method can be executed by a target person's trajectory confirmation device, which can be implemented in the form of hardware and / or software. The target person's trajectory confirmation device can be configured in an electronic device with data processing capabilities. Figure 1 As shown, the method includes:
[0033] S110, determining a confirmed trajectory point and at least two candidate suspected points from the video to be checked according to the original feature information of the target person.
[0034] Among them, the target person may refer to the person whom you want to detect and track. The original feature information can be used to describe the characteristics of the target person. Exemplarily, the original feature information can be represented in the form of feature information, or in the form of image information (image of the target person). The video to be checked may refer to the surveillance video waiting to be checked. The confirmed trajectory point may refer to the confirmed correct trajectory point among the trajectory points of the person in the video to be checked that meet the original feature information. The candidate suspected point may refer to the trajectory point whose correctness cannot be confirmed among the trajectory points of the person in the video to be checked that meet the original feature information.
[0035] In this embodiment, first, based on the original feature information of the target person (which can be provided by the image of the target person), the trajectory points of the person that meet the original feature information can be determined from the video to be checked as candidate trajectory points. Among them, the candidate trajectory points can refer to the trajectory positions of the person that meet the original feature information in the video to be checked, which can be obtained based on the position of the person image acquisition device (such as a camera or a video camera) and the position of the person in the image. Exemplarily, the feature similarity of the original feature information of the target person and the feature information of all persons appearing in the video to be checked are calculated respectively. If the feature similarity is greater than a pre-set similarity threshold, the corresponding person trajectory point is determined as a candidate trajectory point. Then, based on the road network information and the spatiotemporal relationship, the confirmed trajectory points and at least two candidate suspected points are determined from the candidate trajectory points.
[0036] For example, methods such as time overlap and location points being too far apart can be used to first find abnormal trajectory points from candidate trajectory points. Among them, if it is determined that a person cannot reach a certain location based on the chronological order, the candidate trajectory point corresponding to the location is used as an abnormal trajectory point. By excluding abnormal trajectory points, a relatively complete trajectory route can be obtained. This trajectory is mainly confirmed by chronological order. When suspected targets appear at trajectory points with similar time, further investigation is required. Figure 2 This is a schematic diagram of a target person's trajectory provided in the first embodiment of the present invention. Figure 2 As shown in the figure, CAM1, CAM4 and CAM7 are confirmed trajectory points. Suspected target persons appear in CAM2 and CAM3 as well as CAM5 and CAM6, but they cannot meet the requirement of passing by one after another in the time sequence. At this time, CAM2, CAM3 and CAM5, CAM6 can be used as candidate suspected points.
[0037] S120, respectively determining the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, and determining the screening result of the candidate suspected points according to the local small area feature similarity and the feature type of the local small area feature.
[0038] In this embodiment, after determining the confirmed trajectory point and the candidate suspected point, the suspected target personnel in each candidate suspected point and the target personnel in the confirmed trajectory point can be segmented by local features to obtain corresponding local small areas. Exemplary, if the current target is in the front or back state, the head of the current target can be divided into the upper part (forehead area), the middle part (eye area) and the lower part (mouth and nose area), the arm can be divided into the upper arm and the lower arm (the hand is not divided and the feature is not compared), the leg can be divided into the thigh, the calf and the foot, and the body can be divided into 9 small areas according to the nine-square grid. If the current target is in the side state, the head of the current target can be divided into the upper part (above the ear), the middle part (ear) and the lower part (mandible), the arm can be divided into the upper arm and the lower arm (the hand is not divided and the feature is compared), the leg can be divided into the thigh, the calf and the foot, and the body can be divided into 3 small areas of the upper, middle and lower. Among them, since it is a very common situation to hold things with the hands, it can be directly excluded here, and the hands are not divided into small areas and feature comparisons.
[0039] After obtaining the local small area, the feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in each confirmed trajectory point is calculated to obtain the local small area feature similarity. Figure 2 Taking CAM2 in CAM2 as an example, the feature similarities between each small local area of the suspected target person in CAM2 and each corresponding local small area of the target person in CAM1, CAM4 and CAM7 are calculated respectively as the local small area feature similarity of CAM2. It can be understood that for each candidate suspected point, when the local small area of the suspected target person is multiple areas, multiple local small area feature similarities can be obtained after feature similarity calculation, and the number of local small area feature similarities is consistent with the number of local small areas of the suspected target person. After obtaining the local small area feature similarity, the screening result of the candidate suspected point can be determined according to the local small area feature similarity and the feature type of the local small area feature. Among them, the feature type includes temporary features and persistent features. Specifically, temporary features can be understood as features that can be changed at any time, such as a sun hat on the head, an item held in the hand (such as food or a bag) and a sticker on clothes. Persistent features can be understood as fixed features, such as patterns on clothes and tattoos on the body.
[0040] In this embodiment, optionally, before determining the screening results of candidate suspected points based on the similarity of local small area features and the feature type of local small area features, it also includes: determining the feature type of candidate features of the target person based on the comparison results of the original feature information of the target person and the feature information of the target person in the confirmed trajectory points, and / or, determining the feature type of the candidate features of the target person based on the comparison results of the feature information of the target person in the confirmed trajectory points; wherein the feature type includes temporary features and persistent features.
[0041] In this embodiment, the feature type of the candidate feature of the target person can be determined based on the comparison result of the original feature information of the target person and the feature information of the target person in the confirmed trajectory points. Among them, the candidate feature can be a pre-set feature, such as a sun hat on the head and a pattern on the clothes. Specifically, if there is inconsistency between the original feature information of the target person and the candidate feature information of the target person in the confirmed trajectory points, that is, the candidate feature information changes, then it can be determined that the candidate feature information is a temporary feature. In addition, the feature type of the candidate feature of the target person can also be determined based on the comparison result of the feature information of the target person in the confirmed trajectory points. Specifically, if there is inconsistency between the candidate feature information of the target person in the confirmed trajectory points, that is, the candidate feature information changes, then it can be determined that the candidate feature information is a temporary feature.
[0042] In this embodiment, optionally, based on the comparison results of the original feature information of the target person and the feature information of the target person in the confirmed trajectory points, the feature type of the candidate features of the target person is determined, including: based on the comparison results of the feature information of the local small area of the target person in the confirmed trajectory points and the original feature information at the corresponding position, the feature type of the candidate local small area of the target person is determined; correspondingly, based on the comparison results of the feature information of the target person in the confirmed trajectory points, the feature type of the candidate features of the target person is determined, including: based on the comparison results of the feature information of the local small area of the target person in the confirmed trajectory points, the feature type of the candidate local small area of the target person is determined.
[0043] In this embodiment, after obtaining the local small area, the feature information of the candidate local small area can be used as a candidate feature, and then the feature type of the candidate local small area of the target person is determined based on the comparison result of the feature information of the local small area of the target person in the confirmed trajectory point and the original feature information at the corresponding position, and / or, based on the comparison result of the feature information of the local small area of the target person in the confirmed trajectory point. Specifically, if there is a discrepancy between the feature information of the local small area of the target person in the confirmed trajectory point and the original feature information at the corresponding position, it can be determined that the feature type of the candidate local small area is a temporary feature. If there is a discrepancy between the feature information of the local small area of the target person in the confirmed trajectory point, it can be determined that the feature type of the candidate local small area is a temporary feature.
[0044] In this embodiment, optionally, the screening results of the candidate suspected points are determined based on the local small area feature similarity and the feature type of the local small area feature, including: if the local small area feature similarity between any target local small area of the suspected target person in the candidate suspected point and the target local small area of the target person in all confirmed trajectory points is less than the similarity threshold, and the feature type of the target local small area feature is a persistent feature, then the candidate suspected point is determined to be an interference point, and the interference point is screened; if the local small area feature similarity between any target local small area of the suspected target person in the candidate suspected point and the target local small area of the target person in all confirmed trajectory points is greater than or equal to the similarity threshold, and the feature type of the target local small area feature is a persistent feature, then the candidate suspected point is determined to be a confirmed trajectory point; otherwise, the candidate suspected point is retained to obtain the screening results of the candidate suspected points.
[0045] The interference point may refer to any point that is screened out during the screening of suspected points. The similarity threshold may be set according to actual application requirements and is not specifically limited in this embodiment.
[0046] In this embodiment, if the local small area feature similarity between any target local small area of the suspected target person in the candidate suspected point and the target local small area of the target person in all confirmed trajectory points is less than the similarity threshold, and the feature type of the target local small area feature is a persistent feature, then the candidate suspected point is determined to be an interference point, and the interference point is screened; if the local small area feature similarity between any target local small area of the suspected target person in the candidate suspected point and the target local small area of the target person in all confirmed trajectory points is greater than or equal to the similarity threshold, and the feature type of the target local small area feature is a persistent feature, then the candidate suspected point is determined to be a confirmed trajectory point. Except for the above two cases, it is necessary to retain the candidate suspected point to obtain the screening result of the candidate suspected point. That is to say, no matter how the local small area feature similarity between the target local small area of the suspected target person in the candidate suspected point and the target local small area of the target person in the confirmed trajectory point is related to the similarity threshold, as long as the feature type of all local small area features is a temporary feature, the candidate suspected point needs to be retained.
[0047] S130: Determine the trajectory confirmation result of the target person according to the screening result of the candidate suspected points.
[0048] In this embodiment, after determining the screening results of the candidate suspected points, the trajectory confirmation results of the target person can be determined according to the screening results of the candidate suspected points. For example, the candidate suspected points retained in the screening results can be directly used as new confirmed trajectory points, and integrated with the previously determined confirmed trajectory points to obtain the trajectory confirmation results of the target person.
[0049] The technical solution of the embodiment of the present invention determines the confirmed trajectory point and at least two candidate suspected points from the video to be checked based on the original feature information of the target person; respectively determines the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, and determines the screening result of the candidate suspected points based on the local small area feature similarity and the feature type of the local small area feature; determines the trajectory confirmation result of the target person based on the screening result of the candidate suspected points. This technical solution, which checks the suspected points in the video based on the local small area features of the target person and their feature types, can effectively solve the problem of trajectory point confusion in pedestrian re-identification, thereby improving the accuracy of target person trajectory confirmation.
[0050] In this embodiment, optionally, before the interference points are screened, it also includes: expanding the search range of the video to be checked, and after excluding the confirmed trajectory points of the target person, determining the trajectory information of the interfering person in the video to be checked after the expanded range according to the persistent characteristics of the interfering person in the interference points; if the trajectory information of the interfering person meets the independent trajectory condition, the interference points are screened; otherwise, the interference points are retained as suspected points.
[0051] In this embodiment, in order to avoid the situation where the target person's trajectory is incorrectly confirmed due to the error in the screening of suspected points, it is necessary to determine whether the determination of the interference points is reasonable before screening the interference points. Specifically, before screening the interference points, the search range of the video to be checked can be expanded. After excluding the confirmed trajectory points of the target person, the trajectory information of the interference person can be determined in the video to be checked after the expanded scope according to the persistent features of the interference person in the interference points. For example, the persistent features of the interference person in the interference points can be calculated for similarity with the features in the video to be checked after the expanded scope, and the trajectory information of the interference person can be determined based on the similarity calculation result.
[0052] After determining the trajectory information of the interfering person, it is possible to further determine whether the trajectory information of the interfering person meets the independent trajectory condition, and then determine the processing method for the interference point according to the judgment result. Among them, the independent trajectory condition can be understood as a condition that can effectively distinguish the trajectory information of the interfering person from the original trajectory information of the target person. In this embodiment, the independent trajectory condition can be set in advance, and it can be set specifically according to the actual application scenario. For example, the independent trajectory condition can be set so that the trajectory of the interfering person is an obvious independent trajectory, that is, the trajectory information of the interfering person and the original trajectory information of the target person can be effectively distinguished.
[0053] Specifically, the trajectory similarity between the trajectory information of each interfering person and the original trajectory information of the target person is calculated respectively. If the trajectory similarity is less than the preset trajectory similarity threshold, it indicates that the trajectory similarity is low, that is, the trajectory of the interfering person is an obvious independent trajectory, and the trajectory information of the interfering person and the original trajectory information of the target person can be effectively distinguished. At this time, the interference point is determined correctly and the interference point can be directly screened; if the trajectory of the interfering person is short and the trajectory similarity is greater than or equal to the preset trajectory similarity threshold, that is, the trajectory information of the interfering person and the original trajectory information of the target person cannot be effectively distinguished, and no effective new trajectory is formed, and it is still part of the original trajectory information. At this time, it is necessary to retain the interference point as a suspected point and perform manual judgment and confirmation.
[0054] Through such a setting, this scheme can effectively determine whether the interference points are reasonably determined, avoid the situation where the target person's trajectory is incorrectly confirmed due to incorrect screening of suspected points, and further improve the accuracy of the target person's trajectory confirmation.
[0055] Embodiment 2
[0056] Figure 3 A flow chart of a method for confirming the trajectory of a target person provided in the second embodiment of the present invention, this embodiment is optimized based on the above embodiment. The specific optimization is: before respectively determining the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, it also includes: respectively determining the local area feature similarity between the local area of the suspected target person in each candidate suspected point and the local area of the target person in the confirmed trajectory point, and determining the preliminary screening results of the candidate suspected points according to the local area feature similarity as the updated candidate suspected points; if the updated candidate suspected points are not empty, the local areas are divided to obtain local small areas.
[0057] like Figure 3 As shown, the method of this embodiment specifically includes the following steps:
[0058] S210: Determine a confirmed trajectory point and at least two candidate suspected points from the video to be checked according to the original feature information of the target person.
[0059] S220, respectively determine the local area feature similarity between the local area of the suspected target person in each candidate suspected point and the local area of the target person in the confirmed trajectory point, and determine the preliminary screening results of the candidate suspected points according to the local area feature similarity as the updated candidate suspected points.
[0060] In this embodiment, after the confirmed trajectory points and the candidate suspected points are obtained, local feature segmentation can be performed on the suspected target person in each candidate suspected point and the target person in the confirmed trajectory points to obtain corresponding local areas. Figure 4 A schematic diagram of a target local feature segmentation provided in the second embodiment of the present invention. For example, Figure 4 As shown in FIG. 1 , the target can be decomposed into six local regions, namely, the head, hands, feet, and body. After the local regions are determined, the feature similarity between the local region of the suspected target person in each candidate suspected point and the local region of the target person in each confirmed trajectory point is calculated to obtain the local region feature similarity.
[0061] After determining the local area feature similarity, the preliminary screening results of the candidate suspected points can be determined according to the local area feature similarity as the updated candidate suspected points. Optionally, the preliminary screening results of the candidate suspected points are determined according to the local area feature similarity, including: if the local area feature similarities between all local areas of the suspected target person in the candidate suspected points and the local areas of the target person in all confirmed trajectory points are less than the similarity threshold, the candidate suspected points are screened out; otherwise, the candidate suspected points are retained.
[0062] In this embodiment, after obtaining the local area feature similarity, each local area feature similarity can be compared with a pre-set similarity threshold, and the candidate suspected points can be preliminarily screened based on the comparison results. Specifically, if the local area feature similarities between all local areas of the suspected target person in the candidate suspected point and the local areas of the target person in all confirmed trajectory points are less than the similarity threshold, it can be considered that the local features of the candidate suspected point are different from the local features of all confirmed trajectory points, that is, the suspected target person in the candidate suspected point and the target person in the confirmed trajectory point belong to different persons, and the candidate suspected point can be screened out at this time; if there is a local area feature similarity between a local area of the suspected target person in the candidate suspected point and a local area of the target person in a confirmed trajectory point that is greater than or equal to the similarity threshold, it is necessary to further check the candidate suspected point, and the candidate suspected point can be retained at this time.
[0063] Through such a setting, this scheme can preliminarily screen the candidate suspected points based on the size relationship between the local area feature similarity and the similarity threshold, and exclude erroneous suspected points.
[0064] S230: If the updated candidate suspected points are not empty, the local area is divided to obtain small local areas.
[0065] In this embodiment, if the updated candidate suspected point is not empty, it is necessary to further divide the local area into small local areas. Figure 2 For example, assuming that CAM5 and CAM6 are initially screened and CAM5 and CAM6 are retained, CAM5 and CAM6 can be used as updated candidate suspected points. At this time, the updated candidate suspected points are not empty, so it is necessary to divide the local area into small local areas. The determination method of the small local area can refer to the detailed description in the above embodiment 1, which will not be repeated here.
[0066] S240, respectively determine the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, and determine the screening result of the candidate suspected points according to the local small area feature similarity and the feature type of the local small area feature.
[0067] S250: Determine the trajectory confirmation result of the target person according to the screening result of the candidate suspected points.
[0068] The technical solution of the embodiment of the present invention, before respectively determining the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, respectively determine the local area feature similarity between the local area of the suspected target person in each candidate suspected point and the local area of the target person in the confirmed trajectory point, and determine the preliminary screening results of the candidate suspected points according to the local area feature similarity as the updated candidate suspected points; if the updated candidate suspected points are not empty, the local areas are divided to obtain local small areas. This technical solution first performs a preliminary screening of the candidate suspected points based on the local area feature similarity to obtain the updated candidate suspected points. If the updated candidate suspected points are not empty, the final screening results of the candidate suspected points are determined based on the local small area feature similarity and its feature type, which improves the efficiency of the target person trajectory confirmation while ensuring the accuracy of the target person trajectory confirmation.
[0069] Embodiment 3
[0070] Figure 5A flow chart of a method for confirming the trajectory of a target person provided in the third embodiment of the present invention, this embodiment is optimized based on the above embodiment. The specific optimization is: before respectively determining the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, it also includes: determining the direction of the person according to the movement direction of the person in the candidate suspected point or the confirmed trajectory point and the installation direction of the person image acquisition device; correspondingly, respectively determining the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, including: determining the confirmed trajectory point corresponding to each candidate suspected point according to the direction of the person; determining the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the corresponding confirmed trajectory point.
[0071] like Figure 5 As shown, the method of this embodiment specifically includes the following steps:
[0072] S310: Determine a confirmed trajectory point and at least two candidate suspected points from the video to be checked according to the original feature information of the target person.
[0073] S320: Determine the direction of the person according to the moving direction of the person in the candidate suspected point or the confirmed trajectory point and the installation direction of the person image acquisition device.
[0074] In this embodiment, after obtaining the confirmed trajectory points and the candidate suspected points, the direction of the person can be further determined according to the movement direction of the person in the candidate suspected points or confirmed trajectory points and the installation direction of the person image acquisition device. The movement direction of the person can be determined by connecting the position points before and after the person image acquisition device captures the image. Figure 6 A schematic diagram of determining the direction of a person provided by Embodiment 3 of the present invention. is the trajectory vector, used to represent the direction of personnel movement; is the vector directly above the video to be checked. Figure 6 As shown in the figure, since the personnel image acquisition device is usually installed at a certain tilt angle facing forward, by calculating the vector directly above the video to be checked and trajectory vector For example, taking the front view as an example, the angle should be 180 degrees. For the convenience of calculation, it can be considered that the angle between 45 degrees and 135 degrees is the side view, within 45 degrees is the back view, and greater than 135 degrees is the front view. The left and right sides can be to confirm by clicking the direction (from left to right or from right to left).
[0075] S330: Determine a confirmed trajectory point corresponding to each candidate suspected point according to the direction of the person.
[0076] In this embodiment, after determining the direction of the personnel, the personnel direction of each candidate suspected point can be compared with the personnel direction of each confirmed trajectory point to obtain the confirmed trajectory point with the same personnel direction as the candidate suspected point, thereby determining the confirmed trajectory point corresponding to each candidate suspected point.
[0077] S340, determining the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the corresponding confirmed trajectory point.
[0078] For example, Figure 2 Taking CAM2 in the figure as an example, the direction of the person identified by CAM2 is first compared with the directions of the people identified by CAM1, CAM4, and CAM7 respectively. If the directions of the people are consistent, local small area feature comparison is performed. The specific local small area feature comparison method can be found in the detailed description in the above embodiment 1, which will not be repeated here.
[0079] S350, determining the screening result of the candidate suspected points according to the local small area feature similarity and the feature type of the local small area feature.
[0080] S360: Determine the trajectory confirmation result of the target person according to the screening result of the candidate suspected points.
[0081] The technical solution of the embodiment of the present invention, before respectively determining the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, determines the direction of the person according to the movement direction of the person in the candidate suspected point or the confirmed trajectory point and the installation direction of the person image acquisition device; accordingly, determines the confirmed trajectory point corresponding to each candidate suspected point according to the person direction; determines the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the corresponding confirmed trajectory point. This technical solution only compares the local small area features of the candidate suspected points and the confirmed trajectory points with the same personnel direction, thereby improving the accuracy of determining the local small area feature similarity, which helps to further improve the accuracy of target person trajectory confirmation.
[0082] Embodiment 4
[0083] Figure 7 This is a schematic diagram of the structure of a target person's trajectory confirmation device provided in the fourth embodiment of the present invention. The device can execute the target person's trajectory confirmation method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method. Figure 7 As shown, the device comprises:
[0084] The point determination module 410 is used to determine the confirmed trajectory point and at least two candidate suspected points from the video to be checked according to the original feature information of the target person;
[0085] The candidate suspected point screening module 420 is used to respectively determine the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, and determine the screening result of the candidate suspected point according to the local small area feature similarity and the feature type of the local small area feature;
[0086] The trajectory confirmation result determination module 430 is used to determine the trajectory confirmation result of the target person according to the screening result of the candidate suspected points.
[0087] Optionally, the device further comprises:
[0088] A candidate feature type determination module is used to determine the feature type of the candidate feature of the target person based on the comparison result of the original feature information of the target person and the feature information of the target person in the confirmed trajectory point, and / or based on the comparison result of the feature information of the target person in the confirmed trajectory point, before determining the screening result of the candidate suspected point according to the local small area feature similarity and the feature type of the local small area feature; wherein the feature type includes temporary features and persistent features.
[0089] Optionally, the candidate feature type determination module is used to:
[0090] Determine the feature type of the candidate local small area of the target person according to the comparison result of the feature information of the local small area of the target person in the confirmed trajectory point and the original feature information at the corresponding position; and / or,
[0091] According to the comparison result of the feature information of the local small area of the target person in the confirmed trajectory point, the feature type of the candidate local small area of the target person is determined.
[0092] Optionally, the candidate suspected point screening module 420 is used to:
[0093] If the local small area feature similarity between any target local small area of the suspected target person in the candidate suspected point and the target local small area of the target person in all confirmed trajectory points is less than the similarity threshold, and the feature type of the target local small area feature is a persistent feature, then the candidate suspected point is determined to be an interference point, and the interference point is screened;
[0094] If the local small area feature similarity between any target local small area of the suspected target person in the candidate suspected point and the target local small area of the target person in all confirmed trajectory points is greater than or equal to the similarity threshold, and the feature type of the target local small area feature is a persistent feature, then the candidate suspected point is determined to be a confirmed trajectory point;
[0095] Otherwise, the candidate suspected point is retained to obtain a screening result of the candidate suspected point.
[0096] Optionally, the candidate suspected point screening module 420 is further used to:
[0097] Before screening the interference points, the search range of the video to be checked is expanded, and after excluding the confirmed trajectory points of the target person, the trajectory information of the interference person is determined in the expanded video to be checked according to the persistent features of the interference person in the interference points;
[0098] If the trajectory information of the interfering person meets the independent trajectory condition, the interfering point is screened;
[0099] Otherwise, the interference point is retained as a suspected point.
[0100] Optionally, the device further comprises:
[0101] A candidate suspected point updating module is used to determine the local area feature similarity between the local area of the suspected target person in each candidate suspected point and the local area of the target person in the confirmed trajectory point before respectively determining the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, and determine the preliminary screening result of the candidate suspected point according to the local area feature similarity as the updated candidate suspected point;
[0102] The local small area determination module is used to divide the local area to obtain local small areas if the updated candidate suspected point position is not empty.
[0103] Optionally, the device further comprises:
[0104] A personnel direction determination module, used to determine the personnel direction according to the personnel movement direction in the candidate suspected point or the confirmed trajectory point and the installation direction of the personnel image acquisition device before respectively determining the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point;
[0105] Accordingly, the candidate suspected point screening module 420 is used to:
[0106] Determine the confirmed trajectory point corresponding to each candidate suspected point according to the personnel direction;
[0107] Determine the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the corresponding confirmed trajectory point.
[0108] A target person's trajectory confirmation device provided by an embodiment of the present invention can execute a target person's trajectory confirmation method provided by any embodiment of the present invention, and has corresponding functional modules and beneficial effects of the execution method.
[0109] Embodiment 5
[0110] Figure 8 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.
[0111] like Figure 8 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.
[0112] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.
[0113] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a target person's trajectory confirmation method.
[0114] In some embodiments, the target person's trajectory confirmation method can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the target person's trajectory confirmation method described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to execute the target person's trajectory confirmation method by any other appropriate means (e.g., by means of firmware).
[0115] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
[0116] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.
[0117] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0118] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).
[0119] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.
[0120] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.
[0121] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.
[0122] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for confirming the trajectory of a target person, characterized in that: The method comprises: Determine the confirmed trajectory point and at least two candidate suspected points from the video to be checked based on the original feature information of the target person; Determine the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, and determine the screening result of the candidate suspected point according to the local small area feature similarity and the feature type of the local small area feature; The trajectory confirmation result of the target person is determined according to the screening result of the candidate suspected points.
2. The method according to claim 1, characterized in that Before determining the screening result of the candidate suspected points according to the local small area feature similarity and the feature type of the local small area feature, the method further includes: Determine the feature type of the candidate feature of the target person based on the comparison result of the original feature information of the target person and the feature information of the target person in the confirmed trajectory points, and / or based on the comparison result of the feature information of the target person in the confirmed trajectory points; wherein the feature type includes temporary features and permanent features.
3. The method according to claim 2, characterized in that Determining the feature type of the candidate feature of the target person according to the comparison result of the original feature information of the target person and the feature information of the target person in the confirmed trajectory point includes: Determine the feature type of the candidate local small area of the target person according to the comparison result of the feature information of the local small area of the target person in the confirmed trajectory point and the original feature information at the corresponding position; Accordingly, according to the comparison result of the feature information of the target person in the confirmed trajectory point, the feature type of the candidate feature of the target person is determined, including: According to the comparison result of the feature information of the local small area of the target person in the confirmed trajectory point, the feature type of the candidate local small area of the target person is determined.
4. The method according to claim 2 or 3, characterized in that: Determining the screening result of the candidate suspected points according to the local small area feature similarity and the feature type of the local small area feature includes: If the local small area feature similarity between any target local small area of the suspected target person in the candidate suspected point and the target local small area of the target person in all confirmed trajectory points is less than the similarity threshold, and the feature type of the target local small area feature is a persistent feature, then the candidate suspected point is determined to be an interference point, and the interference point is screened; If the local small area feature similarity between any target local small area of the suspected target person in the candidate suspected point and the target local small area of the target person in all confirmed trajectory points is greater than or equal to the similarity threshold, and the feature type of the target local small area feature is a persistent feature, then the candidate suspected point is determined to be a confirmed trajectory point; Otherwise, the candidate suspected point is retained to obtain a screening result of the candidate suspected point.
5. The method according to claim 4, characterized in that Before screening the interference points, the method further includes: Expanding the search range of the video to be checked, after excluding the confirmed trajectory points of the target person, determining the trajectory information of the interfering person in the video to be checked after the expanded range according to the persistent characteristics of the interfering person in the interfering points; If the trajectory information of the interfering person meets the independent trajectory condition, the interfering point is screened; Otherwise, the interference point is retained as a suspected point.
6. The method according to claim 1, characterized in that Before respectively determining the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, the method further includes: Determine the local area feature similarity between the local area of the suspected target person in each candidate suspected point and the local area of the target person in the confirmed trajectory point, and determine the preliminary screening result of the candidate suspected point according to the local area feature similarity as the updated candidate suspected point; If the updated candidate suspected points are not empty, the local area is divided to obtain small local areas.
7. The method according to claim 1, characterized in that Before respectively determining the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, the method further includes: Determine the direction of the person according to the movement direction of the person in the candidate suspected point or confirmed trajectory point and the installation direction of the person image acquisition device; Accordingly, the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point is determined respectively, including: Determine the confirmed trajectory point corresponding to each candidate suspected point according to the personnel direction; Determine the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the corresponding confirmed trajectory point.
8. A target person's trajectory confirmation device, characterized in that: The device comprises: A point determination module is used to determine the confirmed trajectory point and at least two candidate suspected points from the video to be checked according to the original feature information of the target person; A candidate suspected point screening module is used to respectively determine the local small area feature similarity between the local small area of the suspected target person in each candidate suspected point and the local small area of the target person in the confirmed trajectory point, and determine the screening result of the candidate suspected point according to the local small area feature similarity and the feature type of the local small area feature; The trajectory confirmation result determination module is used to determine the trajectory confirmation result of the target person according to the screening result of the candidate suspected points.
9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the target person trajectory confirmation method according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the target person trajectory confirmation method according to any one of claims 1 to 7 when executed.