A method, device, equipment and readable storage medium for determining the position of a target object based on an FDSSST algorithm

Abstract

The present invention provides a method, device, equipment and readable storage medium for determining the position of a target object based on an FDSSST algorithm, and relates to the technical field of target tracking. The method comprises: obtaining first position information of a target object to be tracked in a current image frame and second position information in an image frame previous to the current image frame; determining a first detection confidence of the target object to be tracked in the current image frame according to the first position information; when the first detection confidence does not meet a preset confidence threshold, determining multiple candidate detection points according to the second position information and the size information of the target object to be tracked; and obtaining at least one second detection confidence according to the second position information and the candidate position information corresponding to the multiple candidate detection points; and determining the target position information of the target object to be tracked according to the at least one second detection confidence. The scheme provided by the present invention can improve the accuracy of target tracking, while reducing the probability of algorithm contamination and the amount of calculation.

Description

Technical Field

[0001] The present invention relates to the field of target tracking technology, and in particular to a method, device, equipment and readable storage medium for determining the position of a target object based on an FDSSST algorithm. Background Art

[0002] The fdsst tracking algorithm is a speed-optimized version of the dsst algorithm. It is a commonly used tracking algorithm in target tracking scenarios. It has high detection speed, high accuracy, and can adapt to scene changes and scale changes. It is suitable for embedded real-time tracking systems. When tracking, the fdsst tracking algorithm will continuously train filters based on the existing target shape and surrounding environment in order to adapt to scene changes and scale changes. This causes the target to be considered as the tracking target after occlusion or deformation, and the filter is continuously updated, thereby polluting the filter. After several frames of tracking, the proportion of target parameters after the change in the filter exceeds the proportion of targets before the change. After the target gets rid of occlusion or restores deformation, it will also track the current position instead of the previous target after displacement. Therefore, the fdsst tracking algorithm is often equipped with a template matching or feature matching algorithm to recapture the target when it is lost, but the matching algorithm equipped is generally computationally intensive and time-consuming, and is not suitable for embedded systems. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method, device, equipment and readable storage medium for determining the position of a target object based on the FDST algorithm, so as to improve the accuracy of target object tracking, reduce the amount of tracking calculation, and avoid contamination of the filter based on the FDST algorithm.

[0004] In order to solve the above technical problems, an embodiment of the present invention provides a method for determining the position of a target object based on the FDSSST algorithm, comprising:

[0005] Acquire first position information of a target object to be tracked in a current image frame and second position information of a target object in an image frame previous to the current image frame;

[0006] Determining a first detection confidence of the target object to be tracked in the current image frame according to the first position information;

[0007] Determining whether the first detection confidence satisfies a preset confidence threshold, and when the first detection confidence does not satisfy the preset confidence threshold, determining multiple candidate detection points of the target object to be tracked in the current image frame according to the second position information and the size information of the target object to be tracked;

[0008] Determining at least one second detection confidence of the target object to be tracked according to the second position information and candidate position information corresponding to the plurality of candidate detection points;

[0009] According to at least one of the second detection confidences, target position information of the target object to be tracked in the current image frame is determined.

[0010] Optionally, determining a first detection confidence of the target object to be tracked in the current image frame according to the first position information includes:

[0011] Performing detection processing on the first position information to obtain a first detection result matrix of the target object to be tracked in the current image frame;

[0012] The first detection confidence is obtained according to the first detection result matrix.

[0013] Optionally, judging whether the first detection confidence satisfies a preset confidence threshold, and when the first detection confidence does not satisfy the preset confidence threshold, determining multiple candidate detection points of the target object to be tracked in the current image frame according to the second position information and the size information of the target object to be tracked, including:

[0014] The first detection confidence is compared with the preset confidence threshold. When the first detection confidence is less than the preset confidence threshold, multiple candidate detection points of the target object to be tracked in the current image frame are determined according to the coordinate points corresponding to the second position information and the size information of the target object to be tracked.

[0015] Optionally, determining a plurality of candidate detection points of the target object to be tracked in the current image frame according to the coordinate points corresponding to the second position information and the size information of the target object to be tracked includes:

[0016] Determine the coordinate point corresponding to the second position information as the center point;

[0017] Determine the coordinate adjustment base of the center point according to the size information;

[0018] According to the coordinate adjustment base and the coordinate information of the center point, a plurality of candidate detection points corresponding to the center point are determined.

[0019] Optionally, determining a plurality of candidate detection points corresponding to the center point according to the coordinate adjustment base and the coordinate information of the center point includes:

[0020] According to the coordinate adjustment base and the preset processing magnification, the coordinate information of the center point is linearly processed in a preset orientation to obtain a plurality of candidate detection points corresponding to the center point.

[0021] Optionally, determining at least one second detection confidence of the target object to be tracked according to the second position information and candidate position information corresponding to the plurality of candidate detection points includes:

[0022] sequentially performing detection processing on the second position information and candidate position information corresponding to each candidate detection point in the plurality of candidate detection points to obtain a plurality of second detection result matrices of the target object to be tracked;

[0023] According to the multiple second detection result matrices, the second detection confidence corresponding to each second detection matrix in the multiple second detection result matrices is obtained.

[0024] Optionally, determining target position information of the target object to be tracked in the current image frame according to at least one of the second detection confidences includes:

[0025] In accordance with a preset order, each of at least one of the second detection confidences is compared with the preset confidence threshold in turn; when the current second detection confidence is greater than the preset confidence threshold, the second position information or candidate position information corresponding to the current second detection confidence is determined to be the target position information of the target object to be tracked in the current image frame; otherwise, the next image frame is obtained and the candidate detection points in the next image frame are determined and detected.

[0026] An embodiment of the present invention further provides a target object position determination device based on the fdsst algorithm, comprising:

[0027] An acquisition module, used to acquire first position information of a target object to be tracked in a current image frame and second position information of a target object in an image frame previous to the current image frame;

[0028] A processing module, used to determine a first detection confidence of the target object to be tracked in the current image frame based on the first position information; determine whether the first detection confidence satisfies a preset confidence threshold, and when the first detection confidence does not satisfy the preset confidence threshold, determine multiple candidate detection points of the target object to be tracked in the current image frame based on the second position information and the size information of the target object to be tracked; determine at least one second detection confidence of the target object to be tracked based on the second position information and the candidate position information corresponding to the multiple candidate detection points and the second position information; determine the target position information of the target object to be tracked in the current image frame based on at least one of the second detection confidences.

[0029] An embodiment of the present invention further provides a computing device, comprising: a memory for storing one or more programs; and one or more processors for executing the one or more programs to implement the method described above.

[0030] An embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a program, and when the program is executed by a processor, the method described above is implemented.

[0031] The above solution of the present invention includes at least the following beneficial effects:

[0032] The above solution of the present invention can determine the target position information of the target object to be tracked in the current image frame and ensure the accuracy of target tracking.

[0033] The above scheme of the present invention reduces the contamination of the filter model of the fdsst algorithm in practical scenarios, and prevents the situation in which the algorithm is seriously contaminated when the target undergoes changes such as occlusion, deformation, and exposure, and the changes disappear and the original target cannot be tracked.

[0034] The above scheme of the present invention uses the detection capability of the algorithm itself and implements a target loss recapture strategy at a preset monitoring position by setting a plurality of candidate detection points.

[0035] The above scheme of the present invention, after reducing the probability of contamination of the fdsst algorithm, can realize two major modules of tracking and target loss recapture with one tracking algorithm, thereby reducing the amount of code and process switching between different algorithms, reducing the amount of tracking calculations, and saving calculation time.

[0036] It should be understood that the implementation of any embodiment of the present invention does not mean that multiple or all of the above-mentioned beneficial effects must be possessed or achieved at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0038] The structures, proportions, sizes, etc. illustrated in this specification are only used to match the contents disclosed in the specification so as to facilitate understanding and reading by persons familiar with the technology. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantial technical significance. Any structural modification, change in proportion or adjustment in size shall still fall within the scope of the technical contents disclosed in the present invention without affecting the effects and purposes that can be achieved by the present invention.

[0039] Figure 1 It is a flowchart of a method for determining the position of a target object based on the FDSSST algorithm provided in an embodiment of the present invention.

[0040] Figure 2 is a distribution diagram of a central point and a plurality of candidate detection points provided by an optional embodiment of the present invention;

[0041] Figure 3 is a schematic block diagram of a target object position determination device based on the fdsst algorithm provided by an embodiment of the present invention;

[0042] Figure 4 is a schematic block diagram of an electronic device provided by an embodiment of the present invention; and

[0043] Figure 5 is a schematic block diagram of a computing device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0045] It should be understood that the terms "include / comprise", "consist of..." or any other variations are intended to cover non-exclusive inclusion, so that a product, device, process or method that includes a series of elements includes not only those elements, but also may include other elements not explicitly listed when necessary, or also includes elements inherent to such product, device, process or method. In the absence of more restrictions, the elements defined by the sentence "include / comprise...", "consist of..." do not exclude the presence of other identical elements in the product, device, process or method that includes the elements.

[0046] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0047] like Figure 1 As shown, an embodiment of the present invention proposes a method 10 for determining a target object position based on an FDSSST algorithm, comprising:

[0048] Step 11, obtaining first position information of the target object to be tracked in the current image frame and second position information of the target object in the previous image frame of the current image frame;

[0049] Step 12: determining a first detection confidence of the target object to be tracked in the current image frame according to the first position information;

[0050] Step 13, determining whether the first detection confidence satisfies a preset confidence threshold, and when the first detection confidence does not satisfy the preset confidence threshold, determining multiple candidate detection points of the target object to be tracked in the current image frame according to the second position information and the size information of the target object to be tracked;

[0051] Step 14, determining at least one second detection confidence of the target object to be tracked according to the second position information and the candidate position information corresponding to the plurality of candidate detection points;

[0052] Step 15: Determine target position information of the target object to be tracked in the current image frame according to at least one second detection confidence level.

[0053] In this embodiment, the first position information and the second position information both include coordinate information (horizontal coordinate and vertical coordinate) and position, width and height size information of the target object to be tracked in the corresponding image frame; here, the first position information and the second position information can be obtained by tracking and detecting through the fdsst algorithm;

[0054] The first detection confidence and the second detection confidence both represent the accuracy of tracking detection of the target object to be tracked in the corresponding image frame; based on the first position information, the first detection confidence of the target object to be tracked is obtained to confirm whether the target object to be tracked is accurately tracked in the current image frame, so as to provide a basis for whether to continue tracking and updating the fdsst algorithm filter in the future, and avoid filter pollution caused by continuing tracking and updating the filter with wrong position information;

[0055] Further, the first detection confidence is compared with a preset confidence threshold. When the first detection confidence does not meet the preset confidence threshold, it indicates that the target object to be tracked has failed to be tracked in the current image frame, and the target object to be tracked may be blocked, deformed, exposed, etc.; since the fdsst algorithm needs to update the filter based on the tracked position information, in order to avoid filter contamination caused by continuing to track and update the filter with wrong position information, the fdsst algorithm tracking can be stopped at this time; the target object to be tracked is recaptured and tracked using the second position information in the previous image frame of the current image frame to capture the accurate position information of the target object to be tracked; it should be known that the second position information in the previous image frame should be the correct position information for tracking;

[0056] In this embodiment, if the preset confidence threshold is set too high, the fdsst algorithm will not adapt to changes in the target object to be tracked. If the target object to be tracked changes at this time, it is considered that the tracking has failed, and the filter is almost not contaminated; if the preset confidence threshold is set too low, similar to the previous fdsst algorithm, the tracking is not considered to have failed if the target object to be tracked changes, and the fdsst algorithm will continue to track for multiple frames, resulting in serious filter contamination; therefore, the preset confidence threshold needs to be set according to the current detection environment (the proportion of target occlusion, deformation, or exposure), so that the target changes to a high degree before tracking fails, and no longer continues to track to prevent contamination of the target filter template, thereby retaining the adaptability of the original tracking algorithm and reducing the probability of the filter being contaminated and the risk of tracking the wrong target.

[0057] Here, processing is performed based on the second position information and the size information of the target object to be tracked, and multiple candidate detection points with the point corresponding to the second position information as the center point are determined to expand the recapture and tracking range of the target object to be tracked, and ensure the efficiency and probability of recapture and tracking; in a further step, the center point and multiple candidate detection points are detected and processed to obtain the second detection confidence corresponding to each point. The detection processing here can be performed based on the filter of the existing fdsst algorithm to avoid changing the algorithm twice; the second detection confidence and the preset confidence threshold are used to determine whether the target object to be tracked is successfully recaptured; if the recapture is successful, the point corresponding to the second detection confidence that meets the preset confidence threshold is used as the basis, and the tracking process is re-entered; otherwise, the recapture and tracking detection is continued to determine the accurate target position information;

[0058] When the above-mentioned embodiment of the present invention is applied to an actual tracking system or tracking scene, the C++ version of the fdsst algorithm is first obtained, and a tracking initialization interface, a tracking detection interface, and a tracking release interface are created; wherein the tracking initialization interface is mainly used to set the parameters of the fdsst algorithm, and simultaneously obtain the position and width and height of the reference image and the target object to be tracked, and initialize the algorithm according to the set parameters to obtain the target position related filter and the scale filter; here, the parameters of the fdsst algorithm can be relatively fixed; the tracking detection interface is mainly used to perform tracking detection according to the position information of the target object to be tracked in the last tracking of the acquired image, and judge Check whether the detection confidence meets the preset confidence threshold. If so, continue to run the algorithm and update the position, width, height and two filters of the target object to be tracked. If not, exit this tracking detection and enter the recapture tracking detection process of step 13 and subsequent steps to reduce the contamination of the filter of the fdsst algorithm in practical scenarios, and prevent the fdsst algorithm from being seriously contaminated when the target object to be tracked is occluded, deformed, exposed, etc., and the changes disappear and the original target cannot be tracked. The tracking release interface is mainly used to stop tracking and release the resources requested by the fdsst algorithm to avoid unnecessary waste of resources.

[0059] By determining multiple candidate detection points with the point corresponding to the correct second position information as the center point, and performing recapture and tracking detection based on the filter of the existing fdsst algorithm, a lost recapture strategy for the target object to be tracked can be implemented to ensure the accuracy of recapture and tracking; while reducing the probability of contamination of the fdsst algorithm, based on a tracking algorithm (fdsst algorithm), tracking and recapture tracking after the target object to be tracked is lost can be achieved, thereby reducing the amount of code and the process switching between different algorithms, and further reducing the complexity of tracking.

[0060] In an optional embodiment of the present invention, the above step 12 may include:

[0061] Step 121, performing detection processing on the first position information to obtain a first detection result matrix of the target object to be tracked in the current image frame;

[0062] Step 122: Obtain a first detection confidence level according to the first detection result matrix.

[0063] In this embodiment, the first position information is detected and processed based on the filter in the fdsst algorithm to obtain a first detection result matrix. Furthermore, a first detection confidence can be obtained according to a first indicator value of the first detection result matrix; preferably, the first indicator value may include: a first maximum value, a first average value and a first standard deviation in the first detection result matrix; more preferably, the first detection confidence = (first maximum value - first average value) / first standard deviation; the first position information is detected and processed to obtain the first detection confidence to confirm the accuracy of tracking the target object to be tracked in the current image frame, provide a basis for whether to continue tracking and update the fdsst algorithm filter according to the first position information, avoid filter contamination caused by continuing tracking and updating the filter with erroneous position information, and reduce the probability of contamination.

[0064] Among them, the first detection result matrix obtains the first eigenvalue of the target in the current image frame and the second eigenvalue of the target in the previous image frame by the getFeatures function in the fdsst algorithm, and obtains the feature displacement data by performing feature projection processing on the first eigenvalue and the second eigenvalue, and compares the projection results of the two, and performs Fourier transform on the feature displacement data to obtain res data, which is the result data after feature detection and processing in the fdsst algorithm; the result data is a cv::Mat class (multidimensional array), and then the peak value of the result data is taken as the first maximum value, the mean of the result data is taken as the first average value, and the standard deviation of the result data is taken as the first standard deviation.

[0065] In an optional embodiment of the present invention, the above step 13 may include:

[0066] Step 131, compare the first detection confidence with a preset confidence threshold. When the first detection confidence is less than the preset confidence threshold, determine multiple candidate detection points of the target object to be tracked in the current image frame based on the coordinate points corresponding to the second position information and the size information of the target object to be tracked.

[0067] In this embodiment, the first detection confidence is compared with a preset confidence threshold. When the first detection confidence is less than the preset confidence threshold, it indicates that the position information of the target object to be tracked in the current image frame has failed to be tracked. At this time, the target object to be tracked may be blocked, deformed, exposed, etc. In order to avoid filter contamination caused by continuing to track and update the filter with the wrong first position information, it is necessary to stop tracking, and determine multiple candidate detection points based on the correct second position information of the target object to be tracked in the previous image frame and the size information of the target object to be tracked, so as to achieve the re-capture tracking of the lost target object to be tracked, so as to ensure the correct position information of the lost target object to be tracked is re-captured, and then the integrity, continuity and accuracy of the tracking;

[0068] Preferably, the size information includes: a width value and a height value (or a length value) of the target object to be tracked.

[0069] In an optional embodiment of the present invention, the above step 131 may include:

[0070] Step 1311, determining the coordinate point corresponding to the second position information as the center point;

[0071] Step 1312, determining the coordinate adjustment base of the center point according to the size information;

[0072] Step 1313, determining multiple candidate detection points corresponding to the center point according to the coordinate adjustment cardinality and the coordinate information of the center point.

[0073] In this embodiment, since the probability of the target object to be tracked being lost near the original position is the highest when it is lost, the coordinate point corresponding to the second position information captured correctly in the previous image frame is used as the center point, and multiple candidate detection points near the determined center point are used as the points for loss-recapture detection, so that the probability of loss-recapture can be guaranteed;

[0074] Here, the width value in the size information can be used as the horizontal coordinate adjustment base of the center point, and the height value in the size information can be used as the vertical coordinate adjustment base of the center point; further, the horizontal coordinate value and the vertical coordinate value of the center point are adjusted respectively according to the determined coordinate adjustment base, and the point corresponding to the adjusted coordinate value is determined as the candidate detection point, which provides a detection basis for lost-recapture tracking, and enables lost-recapture tracking to perform lost-recapture detection within the range with the highest loss probability, thereby ensuring the accuracy and probability of lost-recapture tracking.

[0075] In an optional embodiment of the present invention, the above step 1313 may include:

[0076] Step 13131, based on the coordinate adjustment base and the preset processing magnification, linearly process the coordinate information of the center point at the preset orientation to obtain multiple candidate detection points corresponding to the center point.

[0077] In this embodiment, according to the horizontal coordinate adjustment base and the preset processing magnification, the horizontal coordinate value of the center point is linearly processed at the preset orientation to obtain the adjusted horizontal coordinate value; according to the vertical coordinate adjustment base and the preset processing magnification, the vertical coordinate value of the center point is linearly processed at the preset orientation to obtain the adjusted vertical coordinate value; here, the linear processing can be a translation processing at the preset orientation;

[0078] The preset orientation is a number of different orientations set with the center point as the orientation reference, such as Figure 2 As shown, it may include: the upper left position of the center point, the directly above position of the center point, the upper right position of the center point, the directly right position of the center point, the lower right position of the center point, the directly below position of the center point, the lower left position of the center point, and the directly left position of the center point; the center point, the upper left position of the center point, the directly above position of the center point, the upper right position of the center point, the directly right position of the center point, the directly below position of the center point, the lower left position of the center point, and the directly left position of the center point form a nine-square grid structure; wherein the point corresponding to the adjusted coordinate value in each position is a candidate detection point;

[0079] In an implementable example of the present invention, taking the preset position as the upper left position of the center point as an example, the horizontal coordinate value and the vertical coordinate value of the center point are linearly processed at the upper left position of the center point to obtain the processed horizontal coordinate value and vertical coordinate value respectively; specifically, they can be calculated according to the following formula:

[0080] x1=xn×w, y1=yn×h, where x is the horizontal coordinate value of the center point; x1 is the horizontal coordinate value of the candidate detection point corresponding to the upper left position after adjustment; y is the vertical coordinate value of the center point; y1 is the vertical coordinate value of the candidate detection point corresponding to the upper left position after adjustment; n is the preset processing magnification; w is the width value of the target object to be tracked; h is the height value of the target object to be tracked; according to the azimuth distribution of the candidate detection points, and so on, the coordinate value of each candidate detection point can be obtained, specifically:

[0081] The horizontal coordinate value of the candidate detection point corresponding to the directly above position is x2=x; the vertical coordinate value of the candidate detection point corresponding to the directly above position is y2=yn×h;

[0082] The horizontal coordinate value of the candidate detection point corresponding to the upper right position is x3=x+n×w; the vertical coordinate value of the candidate detection point corresponding to the upper right position is y3=yn×h;

[0083] The horizontal coordinate value of the candidate detection point corresponding to the right position is x4=x+n×w; the vertical coordinate value of the candidate detection point corresponding to the right position is y4=y;

[0084] The horizontal coordinate value of the candidate detection point corresponding to the lower right position is x5=x+n×w; the vertical coordinate value of the candidate detection point corresponding to the lower right position is y5=y+n×h;

[0085] The horizontal coordinate value of the candidate detection point corresponding to the position directly below is x6=x; the vertical coordinate value of the candidate detection point corresponding to the position directly below is y6=y+n×h;

[0086] The horizontal coordinate value of the candidate detection point corresponding to the lower left position is x7=xn×w; the vertical coordinate value of the candidate detection point corresponding to the lower left position is y7=y+n×h;

[0087] The horizontal coordinate value of the candidate detection point corresponding to the left position is x8=xn×w; the vertical coordinate value of the candidate detection point corresponding to the left position is y8=y;

[0088] The coordinate values ​​of the center point are translated in different directions to obtain the translated coordinate values, and the points corresponding to the translated coordinate values ​​are used as candidate detection points, and a nine-square grid structure is formed to expand the loss-recapture tracking detection range; since the probability of the target object to be tracked being lost near the original position is the highest, the center point of the correct position information is used as the adjustment and loss-recapture detection benchmark, and nine points (including candidate detection points in eight different preset directions and the center point) are determined to perform loss-recapture tracking detection of the target object to be tracked to ensure the accuracy and probability of loss-recapture.

[0089] In an optional embodiment of the present invention, the above step 14 may include:

[0090] Step 141, sequentially performing detection processing on the second position information and candidate position information corresponding to each candidate detection point in the plurality of candidate detection points, to obtain a plurality of second detection result matrices of the target object to be tracked;

[0091] Step 142: Obtain, according to the plurality of second detection result matrices, a second detection confidence corresponding to each second detection matrix in the plurality of second detection result matrices.

[0092] In this embodiment, the candidate position information corresponding to the candidate detection point is the horizontal coordinate value and the vertical coordinate value of each candidate detection point; based on the filter in the existing fdsst algorithm, the second position information and the candidate position information of each candidate detection point are detected and processed respectively to obtain multiple second detection result matrices; the solution process of the second detection result matrix is ​​the same as that of the above-mentioned first detection result matrix, which will not be repeated here.

[0093] Furthermore, the second detection confidence corresponding to each second detection matrix can be obtained according to the second indicator value of each second detection result matrix; preferably, the second indicator value may include: the second maximum value, the second average value and the second standard deviation in the second detection result matrix; more preferably, the second detection confidence = (second maximum value - second average value) / second standard deviation; the second detection confidence is obtained by detecting and processing the second position information and the candidate position information corresponding to each candidate detection point to confirm whether the lost position information of the target object to be tracked is recaptured to ensure the accuracy of subsequent tracking of the target object to be tracked; at the same time, the existing fdsst algorithm is used in the lost recapture tracking process, and the algorithm is not replaced in the lost recapture tracking process, which does not involve the copying problem of image frames, and also reduces the amount of code and the process switching between different algorithms, thereby reducing the complexity and resource consumption of lost recapture tracking.

[0094] In an optional embodiment of the present invention, the above step 15 may include:

[0095] Step 151, in accordance with a preset order, compare each of the at least one second detection confidence with a preset confidence threshold in turn, and when the current second detection confidence is greater than the preset confidence threshold, determine that the second position information or candidate position information corresponding to the current second detection confidence is the target position information of the target object to be tracked in the current image frame; otherwise, obtain the next image frame and determine and detect the candidate detection points in the next image frame.

[0096] In this embodiment, the preset order may correspond to the preset position, that is, the second detection confidence obtained corresponding to each point position may be compared with the preset confidence threshold in the order of center point position, upper left position of the center point position, directly above position of the center point position, upper right position of the center point position, directly right position of the center point position, lower right position of the center point position, directly below position of the center point position, lower left position of the center point position, and directly left position of the center point, and whether the target object to be tracked is recaptured successfully is determined by the second detection confidence and the preset confidence threshold; when the current second detection confidence is greater than the preset confidence threshold, it is determined that the second position information or candidate position information corresponding to the current second detection confidence is the target position information of the target object to be tracked in the current image frame, then the lost recapture tracking is successful, and further, the second detection confidence that meets the preset confidence threshold is determined. The center point or candidate detection point corresponding to the second detection confidence level is used as the basis, and the tracking process is re-entered; here, when comparing at least one second detection confidence level with a preset confidence threshold, as long as the current second detection confidence level is greater than the preset confidence threshold, it indicates that the lost recapture is successful, and the remaining second detection confidence levels that have not been compared are no longer compared; if all second detection confidence levels are less than the preset confidence threshold after comparison, it indicates that the lost recapture has failed, and each candidate detection point is used as the new center point in a new round of target lost recapture, and a new candidate detection point corresponding to the new center point is set, and the new candidate detection points are cyclically detected in the corresponding order until the recapture is successful, or the recapture wait times out, it is considered that the target is far away from the last tracking point, and it is necessary to enter the search mode to confirm the target position.

[0097] It should be noted that when performing the second detection confidence comparison here, the comparison is performed in the order of: center point, candidate detection point directly above, candidate detection point directly to the right, candidate detection point directly below, candidate detection point directly to the left, candidate detection point on the upper left, candidate detection point on the upper right, candidate detection point on the lower right, and candidate detection point on the lower left.

[0098] The target object position determination method based on the FDSSST algorithm provided by the above-mentioned embodiment of the present invention obtains the first position information of the target object to be tracked in the current image frame and the second position information of the target object to be tracked in the image frame before the current image frame; and determines the first detection confidence of the target object to be tracked in the current image frame according to the first position information; when the first detection confidence does not meet the preset confidence threshold, the tracking step is exited at this time to avoid the filter of the FDSSST algorithm being contaminated in practical scenarios; at the same time, it also avoids the situation that when the target object to be tracked undergoes changes such as occlusion, deformation, exposure, etc., the FDSSST algorithm is seriously contaminated and the changes disappear and the original object cannot be tracked. The situation of the target object to be tracked; further, according to the second position information and the size information of the target object to be tracked, multiple candidate detection points are determined; and according to the second position information and the candidate position information corresponding to the multiple candidate detection points, at least one second detection confidence is obtained; according to the at least one second detection confidence, the target position information of the target object to be tracked is determined to achieve the lost recapture of the target object to be tracked, thereby improving the accuracy of subsequent target tracking; at the same time, when performing lost recapture, detection is performed based on the original fdsst algorithm, and the algorithm is not changed in the middle, which reduces the amount of code and the process switching between different algorithms, and reduces the complexity of lost recapture and resource consumption.

[0099] like Figure 3 As shown, an embodiment of the present invention further provides a target object position determination device 30 based on the fdsst algorithm, comprising:

[0100] An acquisition module 31 is used to acquire first position information of a target object to be tracked in a current image frame and second position information of a target object in an image frame previous to the current image frame;

[0101] The processing module 32 is used to determine a first detection confidence of the target object to be tracked in the current image frame based on the first position information; determine whether the first detection confidence satisfies a preset confidence threshold, and when the first detection confidence does not satisfy the preset confidence threshold, determine multiple candidate detection points of the target object to be tracked in the current image frame based on the second position information and the size information of the target object to be tracked; determine at least one second detection confidence of the target object to be tracked based on the second position information and the candidate position information corresponding to the multiple candidate detection points; and determine the target position information of the target object to be tracked in the current image frame based on the at least one second detection confidence.

[0102] It should be noted that this device is a device corresponding to the above-mentioned method for determining the position of a target object based on the fdsst algorithm. All implementation methods in the above-mentioned method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0103] like Figure 4 As shown, an embodiment of the present invention further provides an electronic device 50, including: a memory 51 for storing one or more computer programs; one or more processors 52 for executing one or more computer programs, and when the computer program is run by the processor, the method as above is executed. All implementations in the above method embodiment are applicable to this embodiment, and the same technical effect can be achieved. The electronic device 50 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices and other similar computing devices. The components shown in the present invention, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required in the present invention.

[0104] like Figure 5 As shown, the electronic device 50 is a computing device, or a computer system, which may include a CPU 501 (computing unit), which can perform various appropriate actions and processes according to a computer program stored in a ROM 502 (read-only memory) or a computer program loaded from a storage unit 508 into a random access RAM 503 (memory). In the RAM 503, various programs and data required for the operation of the device 500 may also be stored. The CPU 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An I / O interface 505 (input / output interface) is also connected to the bus 504.

[0105] A number of components in the electronic device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a disk, an optical disk, etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 allows the device 500 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0106] CPU 501 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of CPU 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. CPU 501 performs the various methods and processes described above. For example, in some embodiments, the target object position determination method 10 based on the fdsst algorithm may be implemented as a computer software program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 508. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 500 via ROM 502 and / or communication unit 509. When the computer program is loaded into RAM 503 and executed by CPU 501, one or more steps of the target object position determination method 10 based on the fdsst algorithm described above may be performed. Alternatively, in other embodiments, the CPU 501 may be configured to execute the target object position determination method 10 based on the fdsst algorithm in any other appropriate manner (for example, by means of firmware).

[0107] The embodiment of the present invention further provides a computer-readable storage medium storing instructions, which, when executed on a computer, enables the computer to execute the above method. All implementations in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0108] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0109] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0110] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0111] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0112] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0113] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, ROM, RAM, disk or optical disk, etc., which can store program codes.

[0114] In addition, it should be pointed out that in the apparatus and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. Moreover, the steps of performing the above series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order, and some steps can be performed in parallel or independently of each other. For those of ordinary skill in the art, it is understandable that all or any steps or components of the method and apparatus of the present invention can be implemented in hardware, firmware, software or a combination thereof in any computing device (including processors, storage media, etc.) or a network of computing devices, which can be achieved by those of ordinary skill in the art using their basic programming skills after reading the description of the present invention.

[0115] Therefore, the purpose of the present invention can also be achieved by running a program or a group of programs on any computing device. The computing device can be a well-known general-purpose device. Therefore, the purpose of the present invention can also be achieved by simply providing a program product containing a program code for implementing a method or device. That is to say, such a program product also constitutes the present invention, and a storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any well-known storage medium or any storage medium developed in the future. It should also be pointed out that in the device and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. In addition, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order. Some steps can be performed in parallel or independently of each other.

[0116] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for determining the position of a target object based on the fdsst algorithm, characterized in that: include: Acquire first position information of a target object to be tracked in a current image frame and second position information of a target object in an image frame previous to the current image frame; Determining a first detection confidence of the target object to be tracked in the current image frame according to the first position information; determining whether the first detection confidence satisfies a preset confidence threshold, and when the first detection confidence does not satisfy the preset confidence threshold, determining, based on the second position information and the size information of the target object to be tracked, a plurality of candidate detection points of the target object to be tracked in the current image frame with the point corresponding to the second position information as the center point, wherein the second position information is the position information correctly tracked in the previous image frame; Wherein, according to the second position information and the size information of the target object to be tracked, determining a plurality of candidate detection points of the target object to be tracked in the current image frame with the point corresponding to the second position information as the center point specifically includes: Determine the coordinate point corresponding to the second position information as the center point; Determine the coordinate adjustment base of the center point according to the size information, wherein the size information includes: the width and height of the target object to be tracked; Determine a plurality of candidate detection points corresponding to the center point according to the coordinate adjustment base and the coordinate information of the center point, wherein the coordinate adjustment base includes taking the width value in the size information as the horizontal coordinate adjustment base of the center point and taking the height value in the size information as the vertical coordinate adjustment base of the center point; and Determine at least one second detection confidence of the target object to be tracked according to the second position information and the candidate position information corresponding to the plurality of candidate detection points, wherein the first detection confidence and the second detection confidence both represent the accuracy of tracking detection of the target object to be tracked in the corresponding image frame, and the first detection confidence and the second detection confidence are both obtained by performing detection processing based on the fdsst algorithm; Determining target position information of the target object to be tracked in the current image frame according to at least one of the second detection confidences specifically includes: In accordance with a preset order, each of at least one of the second detection confidences is compared with the preset confidence threshold in turn; when the current second detection confidence is greater than the preset confidence threshold, the second position information or candidate position information corresponding to the current second detection confidence is determined to be the target position information of the target object to be tracked in the current image frame; otherwise, the next image frame is obtained and the candidate detection points in the next image frame are determined and detected.

2. The method for determining the position of a target object based on the fdsst algorithm according to claim 1, characterized in that: Determining a first detection confidence of the target object to be tracked in the current image frame according to the first position information includes: Performing detection processing on the first position information to obtain a first detection result matrix of the target object to be tracked in the current image frame; The first detection confidence is obtained according to the first detection result matrix.

3. The method for determining the position of a target object based on the fdsst algorithm according to claim 1, characterized in that: Determining a plurality of candidate detection points corresponding to the center point according to the coordinate adjustment base and the coordinate information of the center point, including: According to the coordinate adjustment base and the preset processing magnification, the coordinate information of the center point is linearly processed in a preset orientation to obtain a plurality of candidate detection points corresponding to the center point.

4. The method for determining the position of a target object based on the fdsst algorithm according to claim 1, characterized in that: Determining at least one second detection confidence of the target object to be tracked according to the second position information and candidate position information corresponding to the plurality of candidate detection points, including: Sequentially performing detection processing on the second position information and candidate position information corresponding to each candidate detection point in the plurality of candidate detection points to obtain a plurality of second detection result matrices of the target object to be tracked; According to the multiple second detection result matrices, the second detection confidence corresponding to each second detection matrix in the multiple second detection result matrices is obtained.

5. A device for determining the position of a target object based on the fdsst algorithm, characterized in that: include: An acquisition module, used to acquire first position information of a target object to be tracked in a current image frame and second position information of a target object in an image frame previous to the current image frame; A processing module, configured to determine a first detection confidence of the target object to be tracked in the current image frame according to the first position information; Determine whether the first detection confidence satisfies a preset confidence threshold; when the first detection confidence does not satisfy the preset confidence threshold, determine, based on the second position information and the size information of the target object to be tracked, a plurality of candidate detection points of the target object to be tracked in the current image frame with the point corresponding to the second position information as the center point, wherein the second position information is the position information correctly tracked in the previous image frame; wherein, based on the second position information and the size information of the target object to be tracked, determining a plurality of candidate detection points of the target object to be tracked in the current image frame with the point corresponding to the second position information as the center point specifically includes: Determine the coordinate point corresponding to the second position information as the center point; Determine the coordinate adjustment base of the center point according to the size information, wherein the size information includes: the width and height of the target object to be tracked; Determine a plurality of candidate detection points corresponding to the center point according to the coordinate adjustment base and the coordinate information of the center point, wherein the coordinate adjustment base includes taking the width value in the size information as the horizontal coordinate adjustment base of the center point and taking the height value in the size information as the vertical coordinate adjustment base of the center point; and Determine at least one second detection confidence of the target object to be tracked according to the second position information and the candidate position information corresponding to the plurality of candidate detection points and the second position information, wherein the first detection confidence and the second detection confidence both represent the accuracy of tracking detection of the target object to be tracked in the corresponding image frame, and the first detection confidence and the second detection confidence are both obtained by performing detection processing based on the fdsst algorithm; determine the target position information of the target object to be tracked in the current image frame according to at least one of the second detection confidences, specifically including: In accordance with a preset order, each of at least one of the second detection confidences is compared with the preset confidence threshold in turn; when the current second detection confidence is greater than the preset confidence threshold, the second position information or candidate position information corresponding to the current second detection confidence is determined to be the target position information of the target object to be tracked in the current image frame; otherwise, the next image frame is obtained and the candidate detection points in the next image frame are determined and detected.

6. A computing device, characterized in that include: A memory for storing one or more programs; One or more processors, configured to execute the one or more programs to implement the method according to any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program, which implements the method according to any one of claims 1 to 4 when executed by a processor.

Images