A tracking trajectory reconnection method and device

Abstract

The application provides a tracking trajectory reconnection method and device, wherein the method comprises the following steps: obtaining a first picture sequence corresponding to a first tracking trajectory and a second picture sequence corresponding to a second tracking trajectory; respectively identifying human body features of each pedestrian image data in the first picture sequence and the second picture sequence to obtain a first human body feature set corresponding to the first picture sequence and a second human body feature set corresponding to the second picture sequence, wherein the first human body feature set comprises a plurality of first human body features, and the second human body feature set comprises a plurality of second human body features; calculating a set distance of the first human body feature set and the second human body feature set according to feature distances between each first human body feature and each second human body feature; and connecting the first tracking trajectory and the second tracking trajectory if the set distance is less than a preset threshold. The application can determine whether the first tracking trajectory and the second tracking trajectory can be reconnected, and the obtained determination result has higher reliability.

Description

Technical Field

[0001] This invention relates to the field of intelligent monitoring technology, specifically to a method and apparatus for reconnecting tracking trajectories. Background Technology

[0002] Current trajectory reconnection algorithms select one pedestrian image from each of the two trajectory segments, typically the highest-quality human image from each segment. They then use a human feature extraction model to extract these human features and finally determine whether to reconnect the two trajectories based on the similarity of these features—that is, whether the two trajectories belong to the same pedestrian. However, when the highest-quality human image in a trajectory segment overlaps with other pedestrians, or when the angles and lighting intensities of the two pedestrian images obtained from the two trajectories differ, the similarity of the human features in the two images will be low, making accurate trajectory reconnection impossible. Summary of the Invention

[0003] Therefore, the technical problem to be solved by the present invention is to overcome the defect of the prior art in that it cannot accurately realize trajectory reconnection, thereby providing a tracking trajectory reconnection method and apparatus.

[0004] The first aspect of the present invention provides a method for reconnecting tracking trajectories, comprising: acquiring a first image sequence corresponding to a first tracking trajectory and a second image sequence corresponding to a second tracking trajectory, wherein the first image sequence and the second image sequence respectively contain multiple pedestrian image data; identifying human features of each pedestrian image data in the first image sequence and the second image sequence respectively, to obtain a first human feature set corresponding to the first image sequence and a second human feature set corresponding to the second image sequence, wherein the first human feature set includes multiple first human features and the second human feature set includes multiple second human features; calculating a set distance between the first human feature set and the second human feature set based on the feature distance between each first human feature and each second human feature; and connecting the first tracking trajectory and the second tracking trajectory if the set distance is less than a preset threshold.

[0005] Optionally, in the tracking trajectory reconnection method provided by the present invention, calculating the set distance between the first human feature set and the second human feature set based on the feature distance between each first human feature and each second human feature includes: calculating a first set distance from the first human feature set to the second human feature set and a second set distance from the second human feature set to the first human feature set based on the feature distance between each first human feature and each second human feature; and calculating the set distance between the first human feature set and the second human feature set based on the first set distance and the second set distance.

[0006] Optionally, in the tracking trajectory reconnection method provided by the present invention, calculating the first set distance from the first set of human features to the second set of human features based on the feature distance between each first human feature and each second human feature includes: calculating the distance between each first human feature and the second set of human features respectively, wherein the distance between the first human feature and the second set of human features is the minimum value among the feature distances between the first human feature and each second human feature; and determining the maximum value among the distances between each first human feature and the second set of human features as the first set distance.

[0007] Optionally, in the tracking trajectory reconnection method provided by the present invention, calculating the second set distance from the second set of human features to the first set of human features based on the feature distance between each first human feature and each second human feature includes: calculating the distance between each second human feature and the first set of human features respectively, wherein the distance between the second human feature and the first set of human features is the minimum value among the feature distances between the second human feature and each first human feature; and determining the maximum value among the distances between each second human feature and the first set of human features as the second set distance.

[0008] Optionally, in the tracking trajectory reconnection method provided by the present invention, the feature distance between the first human feature and the second human feature is calculated by the following steps: calculating the similarity between the first human feature and the second human feature based on the vector of the first human feature and the vector of the second human feature; determining the distance between the first human feature and the second human feature based on the similarity, wherein the greater the similarity between the first human feature and the second human feature, the smaller the distance.

[0009] Optionally, in the tracking trajectory reconnection method provided by the present invention, the maximum value of the first set distance and the second set distance is determined as the set distance between the first human feature set and the second human feature set.

[0010] Optionally, in the tracking trajectory reconnection method provided by the present invention, obtaining the first image sequence corresponding to the first tracking trajectory and the second image sequence corresponding to the second tracking trajectory includes: inputting each pedestrian image data corresponding to the first tracking trajectory into a pedestrian image quality detection model to obtain the image quality of each pedestrian image data; obtaining the first image sequence by selecting the set of multiple pedestrian image data with the highest image quality corresponding to the first tracking trajectory; inputting each pedestrian image data corresponding to the second tracking trajectory into the pedestrian image quality detection model to obtain the image quality of each pedestrian image data; and obtaining the second image sequence by selecting the set of multiple pedestrian image data with the highest image quality corresponding to the second tracking trajectory.

[0011] A second aspect of the present invention provides a tracking trajectory reconnection device, comprising: an image acquisition module, configured to acquire a first image sequence corresponding to a first tracking trajectory and a second image sequence corresponding to a second tracking trajectory, wherein the first image sequence and the second image sequence respectively contain multiple pedestrian image data; a human feature recognition module, configured to recognize human features of each pedestrian image data in the first image sequence and the second image sequence respectively, to obtain a first human feature set corresponding to the first image sequence and a second human feature set corresponding to the second image sequence, wherein the first human feature set includes multiple first human features and the second human feature set includes multiple second human features; a feature set distance calculation module, configured to calculate the set distance between the first human feature set and the second human feature set based on the feature distance between each first human feature and each second human feature; and a trajectory reconnection module, configured to connect the first tracking trajectory and the second tracking trajectory if the set distance is less than a preset threshold.

[0012] A third aspect of the present invention provides a computer device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to perform the tracking trajectory reconnection method as provided in the first aspect of the present invention.

[0013] A fourth aspect of the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to perform the tracking trajectory reconnection method provided in the first aspect of the present invention.

[0014] The technical solution of this invention has the following advantages:

[0015] In the tracking trajectory reconnection method and apparatus provided by this invention, a first image sequence corresponding to a first tracking trajectory and a second image sequence corresponding to a second tracking trajectory are first obtained. Both the first and second image sequences contain multiple pedestrian image data. Then, the set distance between the first set of human features and the second set of human features is calculated based on the human features of each pedestrian image data. The first set of human features corresponds to the first tracking trajectory, and the second set of human features corresponds to the second tracking trajectory. Calculating the set distance between the first set of human features and the second set of human features is essentially calculating the similarity between the first and second tracking trajectories. When the set distance is less than a preset threshold, it indicates a high similarity between the first and second tracking trajectories, and the first and second tracking trajectories are then connected. In the tracking trajectory reconnection method and apparatus provided by this invention, the similarity between the first and second tracking trajectories is determined based on multiple pedestrian image data corresponding to the first and second tracking trajectories. Even if one pedestrian image data is abnormal, it will not significantly affect the final calculated set distance. Therefore, by implementing this invention to determine whether the first and second tracking trajectories can be reconnected, the resulting judgment has higher reliability. Attached Figure Description

[0016] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a specific example of the tracking trajectory reconnection method in an embodiment of the present invention;

[0018] Figure 2 This is a schematic diagram illustrating a specific example of the tracking trajectory reconnection device in an embodiment of the present invention.

[0019] Figure 3 This is a schematic diagram illustrating a specific example of a computer device in an embodiment of the present invention. Detailed Implementation

[0020] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] In the description of this invention, it should be noted that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0022] Furthermore, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0023] This invention provides a method for reconnecting tracking trajectories, such as... Figure 1 As shown, it includes:

[0024] Step S11: Obtain the first image sequence corresponding to the first tracking trajectory and the second image sequence corresponding to the second tracking trajectory. The first image sequence and the second image sequence each contain multiple pedestrian image data.

[0025] In one optional embodiment, when tracking and detecting pedestrians, pedestrian image data at different times will be acquired. That is, for each tracking trajectory, there will be multiple different pedestrian image data. Multiple pedestrian image data will be selected from the pedestrian image data corresponding to the first tracking trajectory to form a first image sequence, and multiple pedestrian image data will be selected from the pedestrian image data corresponding to the second tracking trajectory to form a second image sequence.

[0026] In an optional embodiment, the number of pedestrian image data in the first image sequence and the second image sequence is the same. For example, both the first image sequence and the second image sequence include 5 images, the first image sequence = {img1, img2, img3, img4, img5}, and the second image sequence = {img1, img2, img3, img4, img5}.

[0027] Step S12: Identify the human features of each pedestrian image data in the first image sequence and the second image sequence respectively, and obtain the first human feature set corresponding to the first image sequence and the second human feature set corresponding to the second image sequence. The first human feature set includes multiple first human features and the second human feature set includes multiple second human features.

[0028] In an optional embodiment, pedestrian image data from the first image sequence and the second image sequence can be input into a human feature extraction model to obtain human features corresponding to each pedestrian image data. The human features obtained by identifying pedestrian image data from the first image sequence are designated as the first human feature, and the human features obtained by identifying pedestrian image data from the second image sequence are designated as the second human feature. The first and second human features are represented in vector form.

[0029] In one optional embodiment, the human feature extraction model is obtained by training a neural network model. The input of the human feature extraction model is pedestrian image data, and the output is a 512-dimensional vector. The vector output by the human feature extraction model is determined as human features.

[0030] Step S13: Calculate the set distance between the first human feature set and the second human feature set based on the feature distance between each first human feature and each second human feature.

[0031] Determine whether the set distance is less than a preset threshold. If the set distance is less than the preset threshold, determine that the first tracking trajectory and the second tracking trajectory belong to the same target object, and execute step S14.

[0032] Step S14: Connect the first tracking trajectory and the second tracking trajectory.

[0033] If the distance between the sets is greater than or equal to a preset threshold, it is determined that the first tracking trajectory and the second tracking trajectory do not belong to the same pedestrian, and the first tracking trajectory and the second tracking trajectory are not connected.

[0034] In this embodiment of the invention, the smaller the value of the set distance, the higher the similarity between the pedestrian corresponding to the first tracking trajectory and the pedestrian corresponding to the second tracking object.

[0035] In an optional embodiment, the preset threshold can be set according to actual needs; for example, the preset threshold can be set to 0.6.

[0036] In the tracking trajectory reconnection method provided in this embodiment of the invention, a first image sequence corresponding to a first tracking trajectory and a second image sequence corresponding to a second tracking trajectory are first obtained. Both the first and second image sequences contain multiple pedestrian image data. Then, the set distance between the first set of human features and the second set of human features is calculated based on the human features of each pedestrian image data. The first set of human features corresponds to the first tracking trajectory, and the second set of human features corresponds to the second tracking trajectory. Calculating the set distance between the first set of human features and the second set of human features is essentially calculating the similarity between the first and second tracking trajectories. When the set distance is less than a preset threshold, it indicates that the similarity between the first and second tracking trajectories is high, and the first and second tracking trajectories are then connected. In the tracking trajectory reconnection method provided in this embodiment of the invention, the similarity between the first and second tracking trajectories is determined based on multiple pedestrian image data corresponding to the first and second tracking trajectories. Even if one pedestrian image data is abnormal, it will not significantly affect the final calculated set distance. Therefore, the determination result obtained by using this embodiment of the invention to determine whether the first and second tracking trajectories can be reconnected has higher reliability.

[0037] In an optional embodiment, step S13 above, the step of calculating the set distance between the first human feature set and the second human feature set based on the feature distance between each first human feature and each second human feature, specifically includes:

[0038] First, calculate the first set distance from the first human feature set to the second human feature set, and the second set distance from the second human feature set to the first human feature set, based on the feature distance between each first human feature and each second human feature.

[0039] Then, the set distance between the first set of human features and the second set of human features is calculated based on the first set distance and the second set distance.

[0040] In an optional embodiment, the maximum value of the first set distance and the second set distance is determined as the set distance between the first human feature set and the second human feature set.

[0041] In an optional embodiment, the first set distance is calculated through the following steps:

[0042] First, calculate the distance between each first human feature and the second human feature set. The distance between the first human feature and the second human feature set is the minimum value among the feature distances between the first human feature and each second human feature.

[0043] In this embodiment of the invention, for one of the first human body features, the feature distances between the first human body feature and each of the second human body features are calculated, and the minimum value among the multiple feature distances is determined as the distance between the first human body feature and the set of second human body features. This method is used to calculate the distances between each of the first human body features and the set of second human body features.

[0044] Then, the maximum value among the distances between each set of first human features and the set of second human features is determined as the first set distance, i.e., the first set distance. This represents the feature distance between the first human feature and the second human feature.

[0045] In an optional embodiment, the feature distance between the first human feature and the second human feature is calculated through the following steps:

[0046] First, the similarity between the first and second human body features is calculated based on the vectors of the first and second human body features:

[0047]

[0048] Indicates the first human body characteristic With second human characteristics The similarity between the first and second human features is in vector form.

[0049] Then, the distance between the first human feature and the second human feature is determined based on the similarity:

[0050]

[0051] The greater the similarity between the first and second human features, the smaller the distance.

[0052] In an optional embodiment, the second set distance is calculated through the following steps:

[0053] First, calculate the distance between each second human feature and the first human feature set. The distance between the second human feature and the first human feature set is the minimum value among the feature distances between the second human feature and each first human feature.

[0054] In this embodiment of the invention, for one of the second human body features, the feature distances between the second human body feature and each of the first human body features are calculated, and the minimum value among the multiple feature distances is determined as the distance between the second human body feature and the set of first human body features. This method is used to calculate the distances between each second human body feature and the set of first human body features sequentially.

[0055] Then, the maximum value among the distances between each second human feature and the first human feature set is determined as the second set distance, i.e., the second set distance.

[0056] In an optional embodiment, the maximum value between the first set distance and the second set distance is determined as the set distance between the first feature set and the second feature set, i.e., d. H (A,B)=max{d(A,B),d(B,A)}.

[0057] In an optional embodiment, step S11 specifically includes:

[0058] First, the pedestrian image data corresponding to the first tracking trajectory are input into the pedestrian image quality detection model to obtain the image quality of each pedestrian image data. Then, the pedestrian image data corresponding to the second tracking trajectory are input into the pedestrian image quality detection model to obtain the image quality of each pedestrian image data.

[0059] In one optional embodiment, the pedestrian image quality detection model is obtained by training a neural network model. First, a training set and a validation set are acquired. These sets include pedestrian images and corresponding image quality labels for each pedestrian image, where the image quality labels are manually labeled. The training set is then input into the neural network model for iterative training. The accuracy of the trained model is then tested using the validation set. If the model's accuracy reaches a certain value, the currently trained model is determined as the pedestrian image quality detection model.

[0060] Then, the set of multiple pedestrian image data with the highest image quality corresponding to the first tracking trajectory is used to obtain the first image sequence, and the set of multiple pedestrian image data with the highest image quality corresponding to the second tracking trajectory is used to obtain the second image sequence.

[0061] This invention provides a tracking trajectory reconnection device, such as... Figure 2 As shown, it includes:

[0062] The image acquisition module 21 is used to acquire a first image sequence corresponding to the first tracking trajectory and a second image sequence corresponding to the second tracking trajectory. The first image sequence and the second image sequence each contain multiple pedestrian image data. For details, please refer to the description of step S11 in the above embodiment, which will not be repeated here.

[0063] The human body feature recognition module 22 is used to identify the human body features of each pedestrian image data in the first image sequence and the second image sequence respectively, to obtain a first human body feature set corresponding to the first image sequence and a second human body feature set corresponding to the second image sequence. The first human body feature set includes multiple first human body features, and the second human body feature set includes multiple second human body features. For details, please refer to the description of step S12 in the above embodiment, which will not be repeated here.

[0064] The feature set distance calculation module 23 is used to calculate the set distance between the first human feature set and the second human feature set based on the feature distance between each first human feature and each second human feature. For details, please refer to the description of step S13 in the above embodiment, which will not be repeated here.

[0065] The trajectory reconnection module 24 is used to connect the first tracking trajectory and the second tracking trajectory if the set distance is less than a preset threshold. For details, please refer to the description of step S14 in the above embodiment, which will not be repeated here.

[0066] This invention provides a computer device, such as... Figure 3 As shown, the computer device mainly includes one or more processors 31 and a memory 32. Figure 3 Take a processor 31 as an example.

[0067] The computer device may also include an input device 33 and an output device 34.

[0068] The processor 31, memory 32, input device 33, and output device 34 can be connected via a bus or other means. Figure 3 Taking the example of a connection between China and Israel via a bus.

[0069] Processor 31 may be a Central Processing Unit (CPU). Processor 31 may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations thereof. The general-purpose processor may be a microprocessor or any conventional processor. Memory 32 may include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on the use of the tracking trajectory reconnection device. Furthermore, memory 32 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory 32 may optionally include memory remotely located relative to processor 31, and this remote memory may be connected to the tracking trajectory reconnection device via a network. Input device 33 can receive calculation requests (or other numerical or character information) input by the user, and generate key signal inputs related to the tracking trajectory reconnection device. Output device 34 may include a display screen or other display device for outputting calculation results.

[0070] This invention provides a computer-readable storage medium that stores computer instructions. The computer storage medium stores computer-executable instructions that can execute the tracking trajectory reconnection method in any of the above method embodiments. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium may also include combinations of the above types of memory.

[0071] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for reconnecting tracking trajectories, characterized in that, include: Obtain a first image sequence corresponding to a first tracking trajectory and a second image sequence corresponding to a second tracking trajectory, wherein the first image sequence and the second image sequence each contain multiple pedestrian image data; Human features of each pedestrian image data in the first image sequence and the second image sequence are identified respectively to obtain a first human feature set corresponding to the first image sequence and a second human feature set corresponding to the second image sequence. The first human feature set includes multiple first human features and the second human feature set includes multiple second human features. Calculate the set distance between the first human feature set and the second human feature set based on the feature distance between each of the first human features and each of the second human features; If the distance between the sets is less than a preset threshold, the first tracking trajectory and the second tracking trajectory will be connected. Calculate the set distance between the first human feature set and the second human feature set based on the feature distance between each of the first human features and each of the second human features, including: Calculate the first set distance from the first set of human features to the second set of human features, and the second set distance from the second set of human features to the first set of human features, based on the feature distance between each of the first human features and each of the second human features. Calculate the set distance between the first set of human features and the second set of human features based on the first set distance and the second set distance; Calculate the first set distance from the first human feature set to the second human feature set based on the feature distance between each of the first human features and each of the second human features, including: Calculate the distance between each of the first human features and the second human feature set, where the distance between the first human feature and the second human feature set is the minimum value among the feature distances between the first human feature and each of the second human features. The maximum value among the distances between each of the first human features and the second set of human features is determined as the distance of the first set; Calculate the second set distance from the second set of human features to the first set of human features based on the feature distance between each of the first human features and each of the second human features, including: Calculate the distance between each of the second human features and the first human feature set, where the distance between the second human feature and the first human feature set is the minimum value among the feature distances between the second human feature and each of the first human features. The maximum value among the distances between each of the second human features and the first set of human features is determined as the second set distance.

2. The tracking trajectory reconnection method according to claim 1, characterized in that, The feature distance between the first human feature and the second human feature is calculated using the following steps: The similarity between the first human feature and the second human feature is calculated based on the vector of the first human feature and the vector of the second human feature. The distance between the first human feature and the second human feature is determined based on the similarity. The greater the similarity between the first human feature and the second human feature, the smaller the distance.

3. The tracking trajectory reconnection method according to claim 1, characterized in that, The maximum value between the first set distance and the second set distance is determined as the set distance between the first human feature set and the second human feature set.

4. The tracking trajectory reconnection method according to claim 1, characterized in that, Obtain the first image sequence corresponding to the first tracking trajectory and the second image sequence corresponding to the second tracking trajectory, including: The pedestrian image data corresponding to the first tracking trajectory are input into the pedestrian image quality detection model to obtain the image quality of each pedestrian image data; The first image sequence is obtained by taking the set of pedestrian image data with the highest image quality corresponding to the first tracking trajectory; The pedestrian image data corresponding to the second tracking trajectory are input into the pedestrian image quality detection model to obtain the image quality of each pedestrian image data; The second image sequence is obtained by taking the set of pedestrian images with the highest image quality corresponding to the second tracking trajectory.

5. A tracking trajectory reconnection device, characterized in that, include: The image acquisition module is used to acquire a first image sequence corresponding to a first tracking trajectory and a second image sequence corresponding to a second tracking trajectory, wherein the first image sequence and the second image sequence each contain multiple pedestrian image data. The human body feature recognition module is used to identify human body features of each pedestrian image data in the first image sequence and the second image sequence respectively, and obtain a first human body feature set corresponding to the first image sequence and a second human body feature set corresponding to the second image sequence. The first human body feature set includes multiple first human body features, and the second human body feature set includes multiple second human body features. The feature set distance calculation module is used to calculate the set distance between the first human feature set and the second human feature set based on the feature distance between each of the first human features and each of the second human features; The trajectory reconnection module is used to connect the first tracking trajectory and the second tracking trajectory if the set distance is less than a preset threshold. Calculate the set distance between the first human feature set and the second human feature set based on the feature distance between each of the first human features and each of the second human features, including: Calculate the first set distance from the first set of human features to the second set of human features, and the second set distance from the second set of human features to the first set of human features, based on the feature distance between each of the first human features and each of the second human features. Calculate the set distance between the first set of human features and the second set of human features based on the first set distance and the second set distance; Calculate the first set distance from the first human feature set to the second human feature set based on the feature distance between each of the first human features and each of the second human features, including: Calculate the distance between each of the first human features and the second human feature set, where the distance between the first human feature and the second human feature set is the minimum value among the feature distances between the first human feature and each of the second human features. The maximum value among the distances between each of the first human features and the second set of human features is determined as the distance of the first set; Calculate the second set distance from the second set of human features to the first set of human features based on the feature distance between each of the first human features and each of the second human features, including: Calculate the distance between each of the second human features and the first human feature set, where the distance between the second human feature and the first human feature set is the minimum value among the feature distances between the second human feature and each of the first human features. The maximum value among the distances between each of the second human features and the first set of human features is determined as the second set distance.

6. A computer device, characterized in that, include: At least one processor; And a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to perform the tracking trajectory reconnection method as described in any one of claims 1-4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to perform the tracking trajectory reconnection method as described in any one of claims 1-4.

Images