Method for determining tracking object, computer equipment, storage medium and program product

Abstract

The invention relates to the technical field of image processing, and discloses a method for determining a tracking object, computer equipment, a storage medium and a program product, and the method comprises the steps: obtaining a first tracking image of a current frame, a second tracking image of a previous frame, and bounding box information corresponding to a plurality of original tracking objects in the second tracking image; performing image processing on the first tracking image to obtain bounding box information corresponding to a plurality of current tracking objects included in the first tracking image; when it is determined that a plurality of candidate current tracking objects exist according to bounding box information corresponding to the target original tracking object and bounding box information corresponding to each current tracking object, according to the bounding box information of the target original tracking object, the first tracking image, the second tracking image and a pre-constructed image processing model, the current tracking object is tracked; and determining a target current tracking object according to the candidate current tracking objects and the bounding box information corresponding to each candidate current tracking object. The probability of tracking object loss and errors can be reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a method for determining a tracking object, a computer device, a storage medium and a program product. Background Art

[0002] In the fields of video surveillance and intelligent driving, it is generally necessary to track people or cars through tracking algorithms on edge devices. Edge devices refer to computing devices located near the data source or at the edge of the network. They can process data close to where the data is generated, rather than sending all data to a remote data center for processing. However, the computing power resources of edge devices are limited, so the tracking algorithm used by edge devices is generally a Kalman filter algorithm with lower computational complexity.

[0003] In the process of tracking an object, the edge device generally first detects the tracking object from the image and generates relevant information of the tracking object. The tracking information generally includes the category of the tracking object (for example, a person, a car, etc.), the bounding box information (for example, the coordinates of the upper left corner of the bounding box, and the length and width of the bounding box), etc. For each detected tracking object, the edge device can use the Kalman filter algorithm to determine the position information and identification information (ID) of the tracking object in the next frame.

[0004] During the tracking process, the tracking object may be lost or wrong, that is, the same tracking object may have multiple IDs. For example, the ID of tracking object A is "001" and the ID of tracking object B is "002". During the tracking process, due to the close distance between tracking objects A and B, the ID of tracking object A may become "002" and the ID of tracking object B may become "001". Summary of the invention

[0005] In view of this, the present invention provides a method, apparatus, computer device, storage medium and program product for determining a tracking object, so as to reduce the probability of the problem of loss and error of the tracking object.

[0006] In a first aspect, the present invention provides a method for determining a tracking object, the method comprising:

[0007] Acquire a first tracking image of a current frame, a second tracking image of a previous frame corresponding to the current frame, and bounding box information corresponding to a plurality of original tracking objects included in the second tracking image;

[0008] Performing image processing on the first tracking image to obtain bounding box information corresponding to a plurality of currently tracked objects included in the first tracking image;

[0009] Determining whether there is a candidate current tracking object matching the target original tracking object according to the bounding box information corresponding to the target original tracking object and the bounding box information corresponding to each of the current tracking objects, wherein the target original tracking object is any one of the multiple original tracking objects;

[0010] When it is determined that there are multiple candidate current tracking objects, the target current tracking object is determined according to the bounding box information of the target original tracking object, the first tracking image, the second tracking image, the pre-built image processing model, and the bounding box information corresponding to each of the candidate current tracking objects.

[0011] The method for determining a tracking object provided by the present invention has the following advantages:

[0012] This solution can determine the candidate current tracking object that may be the target original tracking object through the bounding box information of the target original tracking object and the bounding box information corresponding to all the current tracking objects. In the case of multiple candidate current tracking objects, the image processing model, the second tracking image where the target original tracking object is located, the first tracking image where the current tracking object is located, and the bounding box information are combined to further process other feature information of the tracking object in the image. In this way, through multi-dimensional matching operations, the target current tracking object that is truly the target original tracking object can be determined. In this way, compared with the method of determining the target current tracking object only through the Kalman filter in the related art, this solution can greatly reduce the probability of tracking object errors and losses.

[0013] In an optional implementation, the bounding box information includes the coordinates of a preset target point on the bounding box, and the length and width of the bounding box;

[0014] The determining whether there is a candidate current tracking object matching the target original tracking object according to the bounding box information corresponding to the target original tracking object and the bounding box information corresponding to each of the current tracking objects includes:

[0015] Determining a target pixel displacement according to coordinates of a target point included in the bounding box information corresponding to the target original tracking object and coordinates of a target point included in the bounding box information corresponding to a first current tracking object, wherein the first current tracking object is any current tracking object among the multiple current tracking objects;

[0016] When it is determined that the target pixel displacement is less than or equal to the preset pixel displacement, the first current tracking object is determined as the candidate current tracking object.

[0017] Specifically, if the position of a tracked object does not change much in two adjacent frames, then the tracked object is likely to be the same target. Therefore, this solution can determine the candidate current tracked object that may be the original target tracked object by calculating the pixel displacement between the original target tracked object and the current tracked object between the two frames and comparing it with the preset pixel displacement.

[0018] In an optional implementation, when it is determined that there are multiple candidate current tracking objects, determining the target current tracking object according to the bounding box information of the target original tracking object, the first tracking image, the second tracking image, the pre-built image processing model, and the bounding box information corresponding to each of the candidate current tracking objects includes:

[0019] intercepting a sub-image corresponding to the target original tracking object from the first tracking image according to the coordinates of the preset target point, the length and the width of the bounding box included in the bounding box information corresponding to the target original tracking object;

[0020] According to the coordinates of the preset target point, the length and the width of the bounding box included in the bounding box information corresponding to the first candidate current tracking object, intercepting a sub-image corresponding to the first candidate current tracking object from the second tracking image, wherein the first candidate current tracking object is any candidate current tracking object among the plurality of candidate current tracking objects;

[0021] The target current tracking object is determined according to the sub-image corresponding to the target original tracking object, the sub-image corresponding to each of the candidate current tracking objects, and the image processing model.

[0022] Specifically, by intercepting sub-images corresponding to the target original tracking object and the candidate current tracking object from the first tracking image and the second tracking image, after intercepting the sub-images, only these small areas need to be processed instead of the entire image. This greatly reduces the computational overhead and improves the processing speed, and is particularly suitable for application in devices with limited resources, such as edge devices. In addition, the sub-images of the target original tracking object and the candidate current tracking object, as well as the pre-built image processing model, are combined for comprehensive analysis and matching. This method takes advantage of the fusion of multiple information and can improve the accuracy of matching.

[0023] In an optional implementation, when the image processing model includes a color model, each of the multiple sub-images includes red, green, and blue (RGB) information corresponding to multiple pixels, respectively, wherein the multiple sub-images include a sub-image corresponding to the target original tracking object and a sub-image corresponding to each of the candidate current tracking objects;

[0024] The step of determining the target current tracking object according to the sub-image corresponding to the target original tracking object, the sub-image corresponding to each of the candidate current tracking objects, and the image processing model comprises:

[0025] Determine the hue information, saturation information and brightness information of each pixel in the first sub-image according to the color model and the RGB information corresponding to each pixel in the first sub-image, wherein the first sub-image is any one of the multiple sub-images;

[0026] Performing statistics on the hue information of each pixel in the first sub-image to determine a main hue range corresponding to the first sub-image;

[0027] Selecting a target pixel from the first sub-image according to a main tone range corresponding to the first sub-image and tone information of each pixel in the first sub-image, wherein the target pixel includes a plurality of pixels;

[0028] After the operation of selecting the target pixel for all sub-images is completed, the color similarity between the target original tracking object and each of the candidate current tracking objects is determined according to the number, hue information, saturation information and brightness information of the target pixels in the sub-image corresponding to the target original tracking object, and the number, hue information, saturation information and brightness information of the target pixels in the sub-image corresponding to each of the candidate current tracking objects;

[0029] When it is determined that there is only one candidate currently tracked object whose color similarity with the target original tracked object is greater than a first preset threshold, the candidate currently tracked object is determined as the target currently tracked object.

[0030] Specifically, by converting each pixel in the sub-image into hue, saturation and brightness information, and statistically analyzing the hue information of each pixel in the sub-image, determining the main hue range, and selecting the target pixel from it, the probability of mismatching due to complex background or partial occlusion is reduced by focusing on the main color features of the tracked object. In addition, this scheme not only considers color information, but also combines the number, hue, saturation and brightness information of pixels, and conducts a multi-dimensional comprehensive analysis, which can improve the reliability of matching.

[0031] In an optional implementation, when it is determined that there are multiple candidate current tracking objects and the color similarity between the target original tracking object is greater than the first preset threshold, the image processing model further includes a contour model, and the method further includes:

[0032] Determine contour feature data corresponding to the target original tracking object according to the hue information, saturation information and brightness information of the target pixel included in the sub-image corresponding to the target original tracking object and the contour model;

[0033] and, determining contour feature data corresponding to the second candidate current tracking object according to the hue information, saturation information and brightness information of the target pixel included in the sub-image corresponding to the second candidate current tracking object, and the contour model, wherein the second candidate current tracking object is any candidate current tracking object among a plurality of candidate current tracking objects whose color similarity with the target original tracking object is greater than the first preset threshold;

[0034] Determining the contour similarity between the target original tracking object and the second candidate current tracking object according to the contour feature data of the target original tracking object and the contour feature data of the second candidate current tracking object;

[0035] When it is determined that there is only one candidate currently tracked object and the contour similarity between the candidate currently tracked object and the target original tracked object is greater than a second preset threshold, the candidate currently tracked object is determined as the target currently tracked object.

[0036] Specifically, when there are multiple candidate currently tracked objects and the color similarity between the target original tracked object is greater than the first preset threshold, further screening is performed through the contour model to determine the target currently tracked object, which is more accurate.

[0037] In an optional implementation, when it is determined that there are multiple candidate currently tracked objects and the contour similarities between the target original tracked object are all greater than the second preset threshold, the method further includes:

[0038] A candidate current tracking object having the greatest contour similarity to the target original tracking object is selected as the target current tracking object.

[0039] Specifically, the object with the highest contour similarity is selected to ensure that the final selected target current tracking object is closest to the target original tracking object in shape and structure.

[0040] In an optional embodiment, the method further includes:

[0041] Acquire identification information of the original tracking object of the target and identification information of the current tracking object of the target;

[0042] Comparing the identification information of the target original tracking object with the identification information of the target current tracking object;

[0043] When it is determined that the identification information of the target original tracking object is different from the identification information of the target current tracking object, the identification information of the target current tracking object is updated to the identification information of the target original tracking object.

[0044] Specifically, since identification information is allocated to each tracked object in the tracking algorithm, this solution can avoid the waste of identification information resources by correcting the tracking information, and can also ensure the accuracy of the tracking information.

[0045] In a second aspect, the present invention provides a device for determining a tracking object, the device comprising:

[0046] an acquisition module, configured to acquire a first tracking image of a current frame, a second tracking image of a previous frame corresponding to the current frame, and bounding box information corresponding to a plurality of original tracking objects included in the second tracking image; perform image processing on the first tracking image to acquire bounding box information corresponding to a plurality of current tracking objects included in the first tracking image;

[0047] A determination module, configured to determine whether there is a candidate current tracking object matching the target original tracking object according to the bounding box information corresponding to the target original tracking object and the bounding box information corresponding to each of the current tracking objects, wherein the target original tracking object is any one of the multiple original tracking objects; when it is determined that there are multiple candidate current tracking objects, determine the target current tracking object according to the bounding box information of the target original tracking object, the first tracking image, the second tracking image, the pre-built image processing model, and the bounding box information corresponding to each of the candidate current tracking objects.

[0048] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the method for determining a tracking object according to the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0049] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the method for determining a tracking object according to the first aspect or any corresponding embodiment thereof.

[0050] In a fifth aspect, the present invention provides a computer program product, comprising computer instructions, wherein the computer instructions are used to enable a computer to execute the method for determining a tracking object according to the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the related technologies, the drawings required for use in the specific embodiments or the related technical descriptions will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0052] Figure 1 is a flow chart of a method for determining a tracking object according to an embodiment of the present invention;

[0053] Figure 2 is a schematic diagram of a process of screening candidate tracking objects by using a color model according to an embodiment of the present invention;

[0054] Figure 3 is a schematic diagram of a process of screening candidate tracking objects by using a color model and a contour model according to an embodiment of the present invention;

[0055] Figure 4 is a schematic diagram of a process of screening candidate current tracking objects through a contour model according to an embodiment of the present invention;

[0056] Figure 5 is a flowchart of a specific example of determining a tracking object according to an embodiment of the present invention;

[0057] Figure 6 is a structural block diagram of an apparatus for determining a tracking object according to an embodiment of the present invention;

[0058] Figure 7 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0060] The present invention can be implemented by a computer device, which can be an edge device. Among them, an edge device refers to a device running at the edge of the network, close to the source of data generation, and capable of local data processing, storage and analysis. For example, it can be a mobile phone, a computer, a vehicle-mounted terminal, etc. By performing tasks at the edge layer of the network, the edge device can reduce the need to transmit data to the data center, thereby improving the response speed and efficiency of data processing. When implementing the tracking function on the edge device, due to the low computing power of the edge device, the selection of the artificial intelligence (AI) algorithm is particularly important.

[0061] The implementation of the tracking function is based on the use of various image processing models and tracking algorithms. Among them, the image processing model may include an object recognition algorithm, a color model, a contour model, etc. For example, the object recognition algorithm may be a YOLOv5 (You Only Look Once version 5) algorithm, and the color model may be a (Hue Saturation Value, HSV) model. The tracking algorithm may be a Kalman filter algorithm.

[0062] The embodiment of the present invention provides a method for determining a tracking object. The image processing model can be used to confirm whether the tracking object determined by the tracking algorithm is accurate, thereby avoiding the problem of incorrect tracking information.

[0063] According to an embodiment of the present invention, a method embodiment for determining a tracking object is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0064] In this embodiment, a method for determining a tracking object is provided, which can be executed by the above-mentioned edge device. Figure 1 is a flow chart of a method for determining a tracking object according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0065] Step S101 : acquiring a first tracking image of a current frame, a second tracking image of a previous frame corresponding to the current frame, and bounding box information corresponding to a plurality of original tracking objects included in the second tracking image.

[0066] The bounding box information may include the coordinates of a preset target point on the bounding box, the length and width of the bounding box, etc. For example, when the bounding box is a rectangle, the preset target point may be the point at the upper left corner of the bounding box.

[0067] Step S102: performing image processing on the first tracking image to obtain boundary box information corresponding to a plurality of currently tracked objects included in the first tracking image.

[0068] Specifically, after receiving the video stream transmitted by the image acquisition device, the edge device can decode the video stream to obtain an image. For each frame of the image, the edge device can perform the following processing:

[0069] First, the identification process of the tracking object.

[0070] The edge device can use the Yolov5 algorithm to identify at least one object and the category of each object from the image. According to the category of each object, it is compared with the preset category list to determine at least one tracking object belonging to the category list. For example, the tracking object can be a person, a car, etc. For each tracking object, the Yolov5 algorithm can output the bounding box information corresponding to each tracking object.

[0071] For each tracked object, the edge device can determine the coordinates of the center point of the bounding box based on the coordinates of the preset target point and the length and width of the bounding box.

[0072] Second, track the prediction of location.

[0073] For each tracked object, the edge device can construct a state quantity corresponding to the tracked object based on the coordinates of the center point, the preset horizontal coordinate change rate and vertical coordinate change rate, and the length and width of the bounding box. The mathematical expression of the state quantity can be as follows:

[0074]

[0075] in, is the state quantity, x is the horizontal coordinate of the center point, y is the vertical coordinate of the center point, Δx is the rate of change of the horizontal coordinate, Δy is the rate of change of the vertical coordinate, w is the width of the bounding box, and h is the length of the bounding box.

[0076] Then, the edge device can determine the target state quantity of the tracked object in the next frame according to the state quantity corresponding to the tracked object, the preset state transfer matrix, the measurement matrix, the preset process noise, the preset measurement error, and the first weight value and the second weight value automatically generated by the tracking algorithm. The specific steps can be as follows:

[0077] In step one, the edge device may determine a first state quantity of the tracked object according to a state transfer matrix, a state quantity corresponding to the tracked object, and a preset process noise.

[0078] For example, step 1 can be expressed as follows:

[0079]

[0080] in, To predict the state quantity, F is the state transfer matrix and w is the preset process noise. For example, the mathematical expression of the state transfer matrix can be as follows:

[0081]

[0082] Wherein, Δt represents the interval between two adjacent frames.

[0083] In step two, the edge device may determine a second state quantity of the tracked object according to a measurement matrix, a state quantity corresponding to the tracked object, and a preset measurement error.

[0084] For example, step 2 can use the following expression:

[0085]

[0086] Among them, z k is the second state quantity, H is the measurement matrix, and v is the preset measurement error. For example, the mathematical expression of the measurement matrix can be as follows:

[0087]

[0088] In step three, the edge device may determine a target state quantity according to the first weight value, the second weight value, the first state quantity, and the second state quantity.

[0089] For example, step three can use the following expression:

[0090]

[0091] Among them, x k+1 is the target state, A is the first weight value, and B is the second weight value.

[0092] Third, determine the tracking object.

[0093] Since the edge device identifies the tracking object in each frame during the tracking process, the displacement of the tracking object between two adjacent frames is relatively small. In this way, in each frame of the image, the edge device can determine which tracking object A in the current frame corresponds to in the previous frame (for example, it can be a tracking object with a state quantity similarity greater than a preset threshold and the highest similarity) based on the first state quantity of the tracking object A constructed in the current frame and the target state quantity of all tracking objects calculated in the previous frame, and assign the identification information of the tracking object corresponding to the previous frame to the tracking object A. If there is no corresponding tracking object in the previous frame, the tracking object A can be determined as a new tracking object identified in the current frame, and a new identification information is assigned to the tracking object A. In this way, in the current frame, the edge device can generate tracking information of the tracking object based on the target state quantity and identification information of the tracking object.

[0094] In this way, during the processing of the current frame, the edge device can first obtain the first tracking image of the current frame, the second tracking image of the previous frame corresponding to the current frame, and the bounding box information and identification information corresponding to the multiple original tracking objects included in the second tracking image (obtained in the processing of the previous frame according to the same processing method as above). The edge device can perform image processing on the first tracking image according to the above process to obtain the bounding box information and identification information corresponding to the multiple current tracking objects included in the first tracking image.

[0095] In the process of determining the tracking object corresponding to the tracking object A in the previous frame, if the tracking object A and the tracking object B are two objects moving side by side, the tracking object may be confirmed incorrectly, resulting in incorrect tracking information. Therefore, in order to avoid the above problem, further matching can be performed through subsequent steps.

[0096] Step S103 : determining whether there is a candidate current tracked object that matches the target original tracked object according to the bounding box information corresponding to the target original tracked object and the bounding box information corresponding to each current tracked object.

[0097] The target original tracking object is any original tracking object among the multiple original tracking objects.

[0098] Specifically, for any tracking object identified in the previous frame, in the processing of the current frame, the edge device can determine whether there is a candidate current tracking object in the following specific manner:

[0099] Step 1: Determine the target pixel displacement according to the coordinates of the target point included in the bounding box information corresponding to the target original tracking object and the coordinates of the target point included in the bounding box information corresponding to the first current tracking object.

[0100] The first current tracking object is any current tracking object among the multiple current tracking objects.

[0101] For example, step 1 can be expressed as follows:

[0102]

[0103] in, is the target pixel displacement, x f is the horizontal coordinate of the preset target point included in the bounding box information corresponding to the first currently tracked object, x i is the horizontal coordinate of the preset target point included in the bounding box information corresponding to the original tracking object, y f is the ordinate of the preset target point included in the bounding box information corresponding to the first currently tracked object, y i It is the ordinate of the preset target point included in the bounding box information corresponding to the target original tracking object.

[0104] Step 2: When it is determined that the target pixel displacement is less than or equal to the preset pixel displacement, the first current tracking object is determined as a candidate current tracking object.

[0105] For example, the preset pixel shift amount may be 20 pixels.

[0106] Specifically, when it is determined that the target pixel displacement is less than or equal to the preset pixel displacement, it means that the first current tracking object may be the target original tracking object, and the first current tracking object can be determined as a candidate current tracking object. When it is determined that the target pixel displacement is greater than the preset pixel displacement, it means that the first current tracking object is not the target original tracking object. At this time, the first current tracking object can be screened out and does not participate in subsequent processing. In this way, it can be determined whether each current tracking object is a candidate tracking object that matches the target original tracking object.

[0107] If there is only one candidate tracking object finally determined, the candidate tracking object can be determined as the target tracking object. If there is no candidate tracking object, no processing is required. If there are multiple candidate tracking objects finally determined, step S104 can be performed to confirm which current tracking object is the target original tracking object.

[0108] Step S104, when it is determined that there are multiple candidate current tracking objects, determine the target current tracking object according to the bounding box information of the target original tracking object, the first tracking image, the second tracking image, the pre-built image processing model, and the bounding box information corresponding to each candidate current tracking object.

[0109] The image processing model may include a color model and / or a contour model. The color model may be an HSV model.

[0110] Specifically, the edge device can perform sub-image capture processing on the first tracking image and the second tracking image respectively, that is, the sub-image corresponding to the target original tracking object can be captured from the first tracking image according to the coordinates of the preset target point included in the bounding box information corresponding to the target original tracking object, the length and width of the bounding box, and the sub-image corresponding to the first candidate current tracking object can be captured from the second tracking image according to the coordinates of the preset target point included in the bounding box information corresponding to the first candidate current tracking object, the length and width of the bounding box. The first candidate current tracking object is any candidate current tracking object among multiple candidate current tracking objects.

[0111] In this way, the edge device can obtain multiple sub-images, including a sub-image corresponding to the target original tracking object and a sub-image corresponding to each candidate current tracking object. The first tracking image and the second tracking image can both be in red green blue (RGB) mode, that is, both tracking images can include RGB information corresponding to multiple pixels, and accordingly, each of the multiple sub-images captured can include RGB information corresponding to multiple pixels.

[0112] Then, the edge device may determine the target current tracking object corresponding to the target original tracking object according to the sub-image corresponding to the target original tracking object, the sub-image corresponding to each candidate current tracking object, and the image processing model.

[0113] Depending on the image processing model, edge devices can perform different processing:

[0114] Case 1: The image processing model only includes the color model. Figure 2 As shown, the processing flow of the screening operation through the color model can be as follows:

[0115] Step 1: Determine hue information, saturation information and brightness information of each pixel in the first sub-image according to the color model and RGB information corresponding to each pixel in the first sub-image.

[0116] The first sub-image is any sub-image among the multiple sub-images.

[0117] Step 2: Count the hue information of each pixel in the first sub-image to determine the main hue range corresponding to the first sub-image.

[0118] Step three: select a target pixel from the first sub-image according to the main tone range corresponding to the first sub-image and the tone information of each pixel in the first sub-image.

[0119] The target pixel includes multiple pixels.

[0120] Step 4, after completing the operation of selecting target pixels for all sub-images, determine the color similarity between the target original tracking object and each candidate current tracking object according to the number, hue information, saturation information and brightness information of the target pixels in the sub-image corresponding to the target original tracking object, and the number, hue information, saturation information and brightness information of the target pixels in the sub-image corresponding to each candidate current tracking object.

[0121] Step 5: When it is determined that there is only one candidate currently tracked object whose color similarity with the target original tracked object is greater than a first preset threshold, the candidate currently tracked object is determined as the target currently tracked object.

[0122] Step 6: When it is determined that the color similarities between multiple candidate current tracking objects and the target original tracking object are all greater than the first preset threshold, the candidate current tracking object with the greatest color similarity to the target original tracking object is selected as the target current tracking object.

[0123] Specifically, when the color model is an HSV model, the edge device can input the first sub-image (that is, the RGB information of all pixels in the first sub-image) into the HSV model to obtain a sub-image converted into the HSV mode. The sub-image in the HSV mode may include HSV information corresponding to multiple pixels, that is, the above-mentioned hue information, saturation information and brightness information. Among them, the hue information can be represented by a color range. For example, [0, 180], the color range can be divided into 36 color sub-ranges, and the size of each color sub-range can be 5. Based on the hue information, the edge device can count the histogram of the HSV-H channel for the first sub-image. The horizontal axis of the histogram can be the color sub-range, and the vertical axis can be the number of pixels in the corresponding color sub-range. In this way, for the first sub-image, the edge device can sort the color sub-ranges from most to least according to the number of pixels in each color sub-range, and determine at least one color sub-range sorted before the preset sort value. Then, the edge device can construct a main hue range based on each color sub-range.

[0124] According to the above processing method, the edge device can determine the main color range corresponding to each sub-image, and the HSV information corresponding to all pixels in each sub-image.

[0125] The following takes the comparison operation between the target original tracking object and the first candidate first tracking object as an example to explain the process of calculating color similarity in detail:

[0126] First, the edge device selects the target pixel in the first main tone range of sub-image A according to the first main tone range corresponding to sub-image A and the tone information corresponding to all pixels in sub-image A. And, according to the second main tone range corresponding to sub-image B and the tone information corresponding to all pixels in sub-image B, the target pixel in the second main tone range of sub-image B is selected. Among them, sub-image A is the sub-image corresponding to the target original tracking object, and sub-image B is the sub-image corresponding to the first candidate current tracking object.

[0127] Then, the edge device can count the number of target pixels in sub-image A and the number of target pixels in sub-image B. Further, the sub-image with the smaller number of target pixels is compared, and the smaller number is determined as the upper limit of the target number. In addition, the edge device can also determine the position of each target pixel in the sub-image to which it belongs. The position can be the row number and column number of the target pixel in the sub-image, and then determine the order of the target pixel in the sub-image to which it belongs according to the position of the target pixel in the sub-image to which it belongs.

[0128] Finally, the edge device can use the following mathematical expression to calculate the color similarity between the target original tracking object and the first candidate current tracking object:

[0129]

[0130] Where d(A,B) represents the color similarity between the original target tracking object and the first candidate current tracking object. V is the brightness information, S is the saturation information, H is the hue information, k is the order of the target pixel in the sub-image to which it belongs, and n is the number of targets. 1k is the target pixel x, x in sub-image A ranked k. 2k is the x, y of the target pixel ranked k in sub-image B 1k is the y of the target pixel ranked k in sub-image A, y 2k is the y, z of the target pixel ranked k in sub-image B 1k is the z of the target pixel ranked k in sub-image A, z 2k is the z of the target pixel ranked k in sub-image B. Among them, x, y, z can be expressed as follows:

[0131]

[0132] In this way, for each candidate current tracking object, the color similarity between it and the target original tracking object can be determined in the above manner. If there is no color similarity between any candidate current tracking object and the target original tracking object greater than the first preset threshold, no processing is required. If there is only one candidate current tracking object and the target original tracking object whose color similarity is greater than the first preset threshold, the candidate current tracking object can be directly determined as the target current tracking object. If there are multiple candidate current tracking objects and the target original tracking object whose color similarity is greater than the first preset threshold, the candidate current tracking object with the largest color similarity can be selected and determined as the target current tracking object.

[0133] In some optional embodiments, the edge device may first determine the overlap between the main color range of the sub-image corresponding to the target original tracking object and the main color range of the sub-image corresponding to the first candidate current tracking object, and determine whether the overlap is greater than a preset overlap threshold. If yes, the color similarity may be further calculated. If not, it may be determined that there is no target current tracking object corresponding to the target original tracking object. Since the main color range can reflect the color similarity between the two images to a certain extent, some candidate current tracking objects may be screened out by firstly using the consistency of the main color range, which may avoid the calculation of color similarity and save more resources.

[0134] Case 2: The image processing model includes a color model and a contour model.

[0135] like Figure 3 As shown, after the edge device completes step 4 in situation 1, if it is determined that the color similarity between multiple candidate current tracking objects and the target original tracking object is greater than the first preset threshold, the edge device can perform contour matching processing on the candidate current tracking objects selected in step 4 of situation 1 according to the contour model. Figure 4 As shown, the processing flow of the screening operation through the contour model can be as follows:

[0136] Step 1: Determine contour feature data corresponding to the target original tracking object based on hue information, saturation information and brightness information of target pixels included in the sub-image corresponding to the target original tracking object and the contour model.

[0137] Step 2: Determine contour feature data corresponding to the second candidate currently tracked object based on hue information, saturation information and brightness information of target pixels included in the sub-image corresponding to the second candidate currently tracked object, and the contour model.

[0138] The second candidate current tracking object is any candidate current tracking object among a plurality of candidate current tracking objects whose color similarity with the target original tracking object is greater than a first preset threshold.

[0139] Step three: determining the contour similarity between the target original tracking object and the second candidate current tracking object according to the contour feature data of the target original tracking object and the contour feature data of the second candidate current tracking object.

[0140] Step 4: when it is determined that there is only one candidate currently tracked object whose contour similarity with the target original tracked object is greater than a second preset threshold, the candidate currently tracked object is determined as the target currently tracked object.

[0141] Step 5: When it is determined that the contour similarities between multiple candidate current tracking objects and the target original tracking object are all greater than the second preset threshold, the candidate current tracking object with the greatest contour similarity to the target original tracking object is selected as the target current tracking object.

[0142] Specifically, the edge device can convert the sub-image of the target original tracking object and the sub-image of the candidate current tracking object screened by the color model (the candidate current tracking object whose color similarity with the target original tracking object is greater than the first preset threshold) from the HSV mode to the RGB mode. During the conversion process, the pixels in the sub-image within the main color range can be converted by the following formulas (10) to (15), and the RGB information of the remaining pixels can be set to preset information, for example, (0, 0, 0). In this way, a sub-image of the RGB mode of the target original tracking object and each candidate current tracking object after color screening can be obtained.

[0143] The data expression corresponding to the RGB information can be as follows:

[0144]

[0145] p=V×(1-S) (13)

[0146] q=V×(1-f×S) (14)

[0147] t=V×(1-(1-f)×S) (15)

[0148] Furthermore, the edge device can convert each sub-image converted into the RGB mode into a grayscale image, and perform binarization processing to obtain the corresponding target sub-image. For each target sub-image, the corresponding contour feature data can be extracted using the contour model (specifically, it can be extracted by polynomial fitting). The contour feature data may include the coordinates of each pixel where the contour is located, that is, the contour feature data is a data set. The edge device can calculate the distance between two data sets according to the Hausdorff Distance formula.

[0149] The Hausdorff distance formula can be as follows:

[0150] d H (A, B) = max (sup a∈A inf b∈B ||ab||, sup b∈B inf a∈A ||ab||) (16)

[0151] Among them, d H (A, B) is the contour similarity between the original target tracking object and the second candidate tracking object, a is a point in data set A, b is a point in data set B, sup represents the supremum, i.e. the maximum value in the data set, and inf represents the infimum, i.e. the minimum value in the data set. ||ab|| is the distance between point a and point b, and the Euclidean distance can be used.

[0152] Case 3: The image processing model only includes the contour model.

[0153] The edge device can extract contour feature data from the sub-image of the target original tracking object and the sub-image of each candidate current tracking object respectively through the contour model, and then, the above-mentioned Hausdorff distance formula, as well as the contour feature data of the target original tracking object and the contour feature data of the first candidate current tracking object can be used to calculate the contour similarity between the target original tracking object and the first candidate current tracking object. And determine whether the contour similarity between the target original tracking object and the first candidate current tracking object is greater than the second preset threshold. When it is determined that there is only one candidate current tracking object whose contour similarity with the target original tracking object is greater than the second preset threshold, the candidate current tracking object is determined as the target current tracking object. When it is determined that there are multiple candidate current tracking objects whose contour similarities with the target original tracking object are all greater than the second preset threshold, the edge device can select the candidate current tracking object with the largest contour similarity with the target original tracking object as the target current tracking object, or further calculate the color similarity between the target original tracking object and the multiple candidate current tracking objects screened by the contour model. When it is determined that there is only one candidate current tracking object whose color similarity with the target original tracking object is greater than the first preset threshold, the edge device can determine the candidate current tracking object as the target current tracking object. When it is determined that there are multiple candidate current tracking objects and the color similarities between the target original tracking object are greater than the first preset threshold, the edge device can select the candidate current tracking object with the greatest color similarity with the target original tracking object as the target current tracking object.

[0154] In some optional implementations, the edge device may compare the identification information of the target original tracking object with the identification information of the target current tracking object. When it is determined that the identification information of the target original tracking object and the identification information of the target current tracking object are different, the identification information of the target current tracking object is updated to the identification information of the target original tracking object.

[0155] Specifically, when the edge device determines that the identification information of the target original tracking object and the identification information of the target current tracking object are the same, it means that the tracking information given by the tracking algorithm is correct and does not need to be modified. When the identification information of the target original tracking object of the edge device and the identification information of the target current tracking object are different, it means that the tracking information given by the tracking algorithm is wrong, and the identification information included in the tracking information of the target current tracking object can be modified to the identification information of the target original tracking object.

[0156] In some optional implementations, in the process of determining the tracking object in step 102, the tracking object identification information may not be assigned first, and accordingly, the edge device may directly add the identification information of the target original tracking object to the tracking information of the target current tracking object. For the current tracking object that does not match any of the original tracking objects, the edge device may determine that it is a newly identified tracking object and may assign new identification information to it.

[0157] In some optional implementations, for each assigned identification information, the edge device can determine whether the identification information appears in the current frame. If not, the value of the counting parameter corresponding to the identification information is increased by one, and it is determined whether the value of the timing parameter is greater than a third preset threshold. If it is greater than the third preset threshold, it can be considered that the tracking object corresponding to the identification information has disappeared and no longer appears. Therefore, the tracking object corresponding to the identification information can be canceled, and the tracking information of the tracking object can be deleted, and the identification information can be recycled. In this way, the storage resources of the edge device can be saved, as well as the resources in the identification information resource pool. In addition, the number of original tracking objects that need to be traversed later can be reduced. Since the edge device needs to perform a large number of matching calculation operations for each original tracking object, timely deletion of related tracking objects can reduce the amount of calculation.

[0158] like Figure 5 As shown in the figure, a specific example diagram of determining a tracking object is given. The edge device can traverse each of the multiple original tracking objects identified in the previous frame. Each time an original tracking object is traversed, it can determine whether there is a corresponding target current tracking object in the multiple current tracking objects identified in the current frame in the above manner. Until all the original tracking objects in the previous frame are traversed. Figure 5 The specific processing of each step has been described in detail in the above content and will not be repeated here.

[0159] The method for determining the tracking object provided in this embodiment is as follows: first, after determining the tracking information, this scheme will further select multiple candidate current tracking objects that may be mistracked through the bounding box information of the target original tracking object and each current tracking object, and further screen out the accurate target current tracking object by calculating the color similarity and contour similarity between the target original tracking object and the candidate current tracking object. This can reduce the probability of loss and error of the tracking object. In addition, this scheme will only perform the next layer of screening through the image processing model when it is determined that there are multiple candidate current tracking objects. In this way, computing power resources can be saved to the maximum extent, and it is suitable for edge devices with low computing power. In addition, in the process of calculating color similarity, only the HSV information of the target pixel within the main color range is considered, which can reduce the amount of calculation.

[0160] In this embodiment, a device for determining a tracking object is also provided, and the device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0161] This embodiment provides a device for determining a tracking object, such as Figure 6 As shown, including:

[0162] The acquisition module 601 is used to acquire a first tracking image of a current frame, a second tracking image of a previous frame corresponding to the current frame, and bounding box information corresponding to a plurality of original tracking objects included in the second tracking image; perform image processing on the first tracking image to acquire bounding box information corresponding to a plurality of current tracking objects included in the first tracking image;

[0163] The determination module 602 is used to determine whether there is a candidate current tracking object matching the target original tracking object according to the bounding box information corresponding to the target original tracking object and the bounding box information corresponding to each current tracking object, wherein the target original tracking object is any one of the multiple original tracking objects; when it is determined that there are multiple candidate current tracking objects, the target current tracking object is determined according to the bounding box information of the target original tracking object, the first tracking image, the second tracking image, the pre-built image processing model, and the bounding box information corresponding to each candidate current tracking object.

[0164] In an optional implementation, the bounding box information includes the coordinates of a preset target point on the bounding box, and the length and width of the bounding box;

[0165] The determination module 602 is specifically configured to:

[0166] Determining a target pixel displacement amount according to coordinates of a target point included in bounding box information corresponding to the target original tracking object and coordinates of a target point included in bounding box information corresponding to a first current tracking object, wherein the first current tracking object is any current tracking object among the multiple current tracking objects;

[0167] When it is determined that the target pixel displacement amount is less than or equal to the preset pixel displacement amount, the first current tracking object is determined as a candidate current tracking object.

[0168] In an optional implementation, the determination module 602 is specifically configured to:

[0169] intercepting a sub-image corresponding to the target original tracking object from the first tracking image according to the coordinates of the preset target point, the length and the width of the bounding box included in the bounding box information corresponding to the target original tracking object;

[0170] According to the coordinates of the preset target point, the length and the width of the bounding box included in the bounding box information corresponding to the first candidate current tracking object, intercepting a sub-image corresponding to the first candidate current tracking object from the second tracking image, wherein the first candidate current tracking object is any candidate current tracking object among the multiple candidate current tracking objects;

[0171] The target current tracking object is determined according to the sub-image corresponding to the target original tracking object, the sub-image corresponding to each candidate current tracking object, and the image processing model.

[0172] In an optional embodiment, when the image processing model includes a color model, each of the multiple sub-images includes red, green, and blue (RGB) information corresponding to multiple pixels, respectively, wherein the multiple sub-images include a sub-image corresponding to the target original tracking object and a sub-image corresponding to each candidate current tracking object;

[0173] The determination module 602 is specifically configured to:

[0174] Determine hue information, saturation information, and brightness information of each pixel in the first sub-image according to the color model and RGB information corresponding to each pixel in the first sub-image, wherein the first sub-image is any one of the multiple sub-images;

[0175] Collecting statistics on the hue information of each pixel in the first sub-image to determine a main hue range corresponding to the first sub-image;

[0176] Selecting a target pixel from the first sub-image according to a main tone range corresponding to the first sub-image and tone information of each pixel in the first sub-image, wherein the target pixel includes a plurality of pixels;

[0177] After the operation of selecting target pixels for all sub-images is completed, the color similarity between the target original tracking object and each candidate current tracking object is determined according to the number, hue information, saturation information and brightness information of the target pixels in the sub-image corresponding to the target original tracking object, and the number, hue information, saturation information and brightness information of the target pixels in the sub-image corresponding to each candidate current tracking object;

[0178] When it is determined that there is only one candidate currently tracked object whose color similarity with the target original tracked object is greater than the first preset threshold, the candidate currently tracked object is determined as the target currently tracked object.

[0179] In an optional implementation, when it is determined that the color similarities between multiple candidate current tracking objects and the target original tracking object are all greater than the first preset threshold, the image processing model further includes a contour model, and the determination module 602 is specifically used to:

[0180] Determine contour feature data corresponding to the target original tracking object according to hue information, saturation information and brightness information of target pixels included in the sub-image corresponding to the target original tracking object, and the contour model;

[0181] and, determining contour feature data corresponding to the second candidate current tracking object according to hue information, saturation information and brightness information of target pixels included in the sub-image corresponding to the second candidate current tracking object, and the contour model, wherein the second candidate current tracking object is any candidate current tracking object among a plurality of candidate current tracking objects whose color similarity with the target original tracking object is greater than a first preset threshold;

[0182] Determine the contour similarity between the target original tracking object and the second candidate current tracking object according to the contour feature data of the target original tracking object and the contour feature data of the second candidate current tracking object;

[0183] When it is determined that there is only one candidate currently tracked object and the contour similarity between the candidate currently tracked object and the target original tracked object is greater than the second preset threshold, the candidate currently tracked object is determined as the target currently tracked object.

[0184] In an optional implementation, the determination module 602 is specifically configured to:

[0185] When it is determined that there are multiple candidate current tracking objects whose contour similarities with the target original tracking object are all greater than the second preset threshold, the candidate current tracking object with the greatest contour similarity with the target original tracking object is selected as the target current tracking object.

[0186] In an optional implementation, the device further includes an updating module:

[0187] Obtaining identification information of the original tracking object of the target and identification information of the current tracking object of the target;

[0188] Compare the identification information of the original tracking object of the target with the identification information of the current tracking object of the target;

[0189] When it is determined that the identification information of the target original tracking object is different from the identification information of the target current tracking object, the identification information of the target current tracking object is updated to the identification information of the target original tracking object.

[0190] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0191] The device for determining the tracking object in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0192] The embodiment of the present invention also provides a computer device having the above Figure 6 The device for determining a tracking object is shown.

[0193] See also Figure 7 , Figure 7 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Figure 7 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 7 A processor 10 is taken as an example.

[0194] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware integrated circuit. The hardware integrated circuit may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable logic gate array, a general purpose array logic or any combination thereof.

[0195] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0196] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0197] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0198] The computer device further comprises a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0199] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0200] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.

[0201] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for determining a tracking object, characterized in that: The method comprises: Acquire a first tracking image of a current frame, a second tracking image of a previous frame corresponding to the current frame, and bounding box information corresponding to a plurality of original tracking objects included in the second tracking image; Performing image processing on the first tracking image to obtain bounding box information corresponding to a plurality of currently tracked objects included in the first tracking image; Determining whether there is a candidate current tracking object matching the target original tracking object according to the bounding box information corresponding to the target original tracking object and the bounding box information corresponding to each of the current tracking objects, wherein the target original tracking object is any one of the multiple original tracking objects; When it is determined that there are multiple candidate current tracking objects, the target current tracking object is determined according to the bounding box information of the target original tracking object, the first tracking image, the second tracking image, the pre-built image processing model, and the bounding box information corresponding to each of the candidate current tracking objects.

2. The method according to claim 1, characterized in that The bounding box information includes the coordinates of a preset target point on the bounding box, and the length and width of the bounding box; The determining whether there is a candidate current tracking object matching the target original tracking object according to the bounding box information corresponding to the target original tracking object and the bounding box information corresponding to each of the current tracking objects includes: Determining a target pixel displacement according to coordinates of a target point included in the bounding box information corresponding to the target original tracking object and coordinates of a target point included in the bounding box information corresponding to a first current tracking object, wherein the first current tracking object is any current tracking object among the multiple current tracking objects; When it is determined that the target pixel displacement is less than or equal to the preset pixel displacement, the first current tracking object is determined as the candidate current tracking object.

3. The method according to claim 2, characterized in that When it is determined that there are a plurality of candidate current tracking objects, determining the target current tracking object according to the bounding box information of the target original tracking object, the first tracking image, the second tracking image, the pre-built image processing model, and the bounding box information corresponding to each of the candidate current tracking objects, comprises: intercepting a sub-image corresponding to the target original tracking object from the first tracking image according to the coordinates of the preset target point, the length and the width of the bounding box included in the bounding box information corresponding to the target original tracking object; According to the coordinates of the preset target point, the length and the width of the bounding box included in the bounding box information corresponding to the first candidate current tracking object, intercepting a sub-image corresponding to the first candidate current tracking object from the second tracking image, wherein the first candidate current tracking object is any candidate current tracking object among the plurality of candidate current tracking objects; The target current tracking object is determined according to the sub-image corresponding to the target original tracking object, the sub-image corresponding to each of the candidate current tracking objects, and the image processing model.

4. The method according to claim 3, characterized in that When the image processing model includes a color model, each of the multiple sub-images includes red, green, and blue (RGB) information corresponding to multiple pixels, wherein the multiple sub-images include a sub-image corresponding to the target original tracking object and a sub-image corresponding to each of the candidate current tracking objects; The step of determining the target current tracking object according to the sub-image corresponding to the target original tracking object, the sub-image corresponding to each of the candidate current tracking objects, and the image processing model comprises: Determine the hue information, saturation information and brightness information of each pixel in the first sub-image according to the color model and the RGB information corresponding to each pixel in the first sub-image, wherein the first sub-image is any one of the multiple sub-images; Performing statistics on the hue information of each pixel in the first sub-image to determine a main hue range corresponding to the first sub-image; Selecting a target pixel from the first sub-image according to a main tone range corresponding to the first sub-image and tone information of each pixel in the first sub-image, wherein the target pixel includes a plurality of pixels; After the operation of selecting the target pixel for all sub-images is completed, the color similarity between the target original tracking object and each of the candidate current tracking objects is determined according to the number, hue information, saturation information and brightness information of the target pixels in the sub-image corresponding to the target original tracking object, and the number, hue information, saturation information and brightness information of the target pixels in the sub-image corresponding to each of the candidate current tracking objects; When it is determined that there is only one candidate currently tracked object whose color similarity with the target original tracked object is greater than a first preset threshold, the candidate currently tracked object is determined as the target currently tracked object.

5. The method according to claim 4, characterized in that When it is determined that there are a plurality of candidate current tracking objects and the target original tracking object, the color similarities of which are all greater than the first preset threshold, the image processing model further includes a contour model, and the method further includes: Determine contour feature data corresponding to the target original tracking object according to the hue information, saturation information and brightness information of the target pixel included in the sub-image corresponding to the target original tracking object and the contour model; and, determining contour feature data corresponding to the second candidate current tracking object according to the hue information, saturation information and brightness information of the target pixel included in the sub-image corresponding to the second candidate current tracking object, and the contour model, wherein the second candidate current tracking object is any candidate current tracking object among a plurality of candidate current tracking objects whose color similarity with the target original tracking object is greater than the first preset threshold; Determining the contour similarity between the target original tracking object and the second candidate current tracking object according to the contour feature data of the target original tracking object and the contour feature data of the second candidate current tracking object; When it is determined that there is only one candidate currently tracked object and the contour similarity between the candidate currently tracked object and the target original tracked object is greater than a second preset threshold, the candidate currently tracked object is determined as the target currently tracked object.

6. The method according to claim 5, characterized in that When it is determined that there are multiple candidate currently tracked objects and the target original tracked object whose contour similarities are all greater than the second preset threshold, the method further includes: A candidate current tracking object having the greatest contour similarity to the target original tracking object is selected as the target current tracking object.

7. The method according to any one of claims 1 to 6, characterized in that: The method further comprises: Acquire identification information of the original tracking object of the target and identification information of the current tracking object of the target; Comparing the identification information of the target original tracking object with the identification information of the target current tracking object; When it is determined that the identification information of the target original tracking object is different from the identification information of the target current tracking object, the identification information of the target current tracking object is updated to the identification information of the target original tracking object.

8. A device for determining a tracking object, characterized in that: The device comprises: an acquisition module, configured to acquire a first tracking image of a current frame, a second tracking image of a previous frame corresponding to the current frame, and bounding box information corresponding to a plurality of original tracking objects included in the second tracking image; perform image processing on the first tracking image to acquire bounding box information corresponding to a plurality of current tracking objects included in the first tracking image; A determination module, configured to determine whether there is a candidate current tracking object matching the target original tracking object according to the bounding box information corresponding to the target original tracking object and the bounding box information corresponding to each of the current tracking objects, wherein the target original tracking object is any one of the multiple original tracking objects; when it is determined that there are multiple candidate current tracking objects, determine the target current tracking object according to the bounding box information of the target original tracking object, the first tracking image, the second tracking image, the pre-built image processing model, and the bounding box information corresponding to each of the candidate current tracking objects.

9. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method for determining a tracking object according to any one of claims 1 to 7 by executing the computer instructions.

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 computer to execute the method for determining a tracking object according to any one of claims 1 to 7.

Images