A method, device and medium for real-time posture detection of small targets in dense crowds

Abstract

The application relates to the field of machine vision, and particularly discloses a method and device for detecting the real-time posture of a small target in a dense crowd and a medium, which adopts a modified Alphapose framework to build a real-time posture detection model, specifically includes replacing the end-side target detection algorithm in the Alphapose framework with a YOLOV10 algorithm; setting the human body detection confidence threshold in the YOLOV10 algorithm to a minimum value; adopting a non-maximum suppression algorithm based on a CIOU loss function to post-process the YOLOV10 algorithm; and performing skeleton point detection on n human body detection frames pointing to the same target through a Fastpose skeleton point detection network. The application realizes the detection of the posture of a small target by improving the human body detector and the post-processing logic of the skeleton point detection, and has high detection efficiency and is not prone to losing the target.

Description

Technical Field

[0001] The present invention relates to the field of machine vision, and in particular to a method, device and medium for real-time posture detection of small targets in a dense crowd. Background Art

[0002] With the rapid development of machine vision, the diversity of various human detection algorithms is increasing, and the efficiency and accuracy of recognizing human targets, expressions, postures, and other data are becoming increasingly higher. Currently, human posture detection mainly involves two directions: end-to-end posture detection, and two-stage posture detection; the two stages are human body detection and human skeleton point detection. However, in actual project applications, it is found that existing human posture detection algorithms still have some unresolved issues, especially the Alphapose algorithm. Although its recognition accuracy for multi-person posture estimation in natural scenes with dense crowds is higher than that of other similar models, it consumes a lot of computing resources and has a slow detection speed. Real-time detection of the real-time posture of small targets in dense crowds requires extremely advanced computing platform support.

[0003] The end-to-end pose detection neural network consists of a human detection network branch and a skeleton point detection branch. Compared to two-stage pose detection, end-to-end pose detection is constrained by both branches, resulting in poor generalization for small objects. Furthermore, the accuracy of the two-stage pose detection algorithm is significantly affected by the human detection results.

[0004] The original Alphapose algorithm (multi-person pose estimation algorithm) uses the YOLOX human detector or the YOLOV3-SPP human detector. The network parameters of the above human detectors are relatively large. Although the recall rate of small targets has been improved, the inference time has also increased. Although there are attempts to replace other detectors to improve detection efficiency in the existing technology, the detection efficiency is still not ideal, and the calculation of conventional hardware is still difficult to meet sufficient recognition efficiency. In actual scenarios, since the human body detection frame does not necessarily correspond to a complete human body, it is easy to cause the human skeleton point detection to lose some skeleton points, resulting in inaccurate final detection results. Summary of the Invention

[0005] In order to overcome the above problems, the present invention provides a method, device and medium for real-time posture detection of small targets in a dense crowd.

[0006] The present invention provides a method for real-time posture detection of small targets in dense crowds, comprising: constructing a real-time posture detection model using a modified Alphapose framework;

[0007] Use multiple pose-labeled images and divide them into training set, validation set and test set;

[0008] The real-time posture detection model is trained using the training set and the validation set until the real-time posture detection model passes the test of the test set;

[0009] Use the tested real-time posture detection model to perform real-time posture detection on small targets in dense crowds and output real-time posture detection results;

[0010] The modified Alphapose framework is specifically:

[0011] Use the YOLO V10 algorithm to replace the on-device object detection algorithm in the Alphapose framework to obtain human detection frames in the image.

[0012] Set the human detection confidence threshold in the YOLOV10 algorithm to the minimum value;

[0013] The non-maximum suppression algorithm based on the CIOU loss function is used for post-processing of the YOLOV10 algorithm to screen n human detection boxes pointing to the same target;

[0014] Through the Fastpose skeleton point detection network, skeleton point detection is performed on n human detection boxes pointing to the same target, and n skeleton point detection results of the target are extracted;

[0015] Traverse the 26 skeleton points corresponding to the n skeleton point detection results, take the skeleton point with the largest confidence, and reassemble it into the posture detection result corresponding to the 26 human skeleton points.

[0016] Preferably, in the non-maximum suppression algorithm based on the CIOU loss function, the calculation formula of the CIOU loss function is as follows:

[0017]

[0018] Among them, w' and h' are the width and height of the intersection of the two human detection frames, w1, h1, w2 and h2 are the width and height of the two human detection frames respectively, ρ is the distance between the center points of the two human detection frames, and c is the length of the hypotenuse of the bounding box composed of the two targets.

[0019] Preferably, the YOLOV10 algorithm specifically adopts one of the YOLOV10n algorithm or the YOLOV10s algorithm.

[0020] Preferably, the non-maximum suppression algorithm based on the CIOU loss function is used to post-process the YOLOV10 algorithm to screen n human detection frames pointing to the same target, specifically including the following steps:

[0021] Step a1: export all human detection boxes pointing to the same target in the human detection boxes identified by the YOLOV10 algorithm with the human detection confidence threshold set to the minimum value as a list BoxLists, and create a new list ResultBoxLists for saving the final results;

[0022] Step a2: sort each human detection box in the list BoxLists in descending order of confidence to obtain the list SortBoxLists;

[0023] Step a3: traverse each human detection frame in the list SortBoxLists, calculate the CIOU loss function for the first human detection frame ranked first in the list SortBoxLists and the remaining human detection frames, and record the human detection frames whose CIOU loss function calculation results are greater than 0.25;

[0024] Step a4: insert the human body detection box ranked first in the SortBoxLists list into the ResultBoxLists list;

[0025] Step a5, deleting the first human detection frame in the list SortBoxLists and the human detection frame recorded in step a3 from the list SortBoxLists;

[0026] Step a6, looping through steps a3 to a5 until the number of human body detection boxes in the list SortBoxLists is less than or equal to one;

[0027] Step a7: insert the remaining human detection frames in the list SortBoxLists into the list ResultBoxLists, and output n human detection frames pointing to the target; where n is the number of human detection frames finally included in the list ResultBoxLists.

[0028] Preferably, the post-processing of the YOLOV10 algorithm using a non-maximum suppression algorithm based on a CIOU loss function to screen n human detection frames pointing to the same target also includes the following steps:

[0029] In the human detection frame identified by the YOLOV10 algorithm with the human detection confidence threshold set to the minimum value, all human detection frames pointing to the same target are calculated, and the IOU loss function is calculated. The human detection frames with a calculation result greater than 0.8 are merged into a new human detection frame.

[0030] The present invention also provides a device for real-time posture detection of small targets in dense crowds, comprising: a building module, a training module, a testing module and a detection module;

[0031] The establishing module is used for constructing a real-time posture detection model by using the modified Alphapose framework.

[0032] The training module is used for adopting multiple posture-labeled images, which are divided into a training set, a verification set and a test set.

[0033] The testing module is used for training the real-time posture detection model by using the training set and the verification set until the real-time posture detection model passes the test of the test set.

[0034] The detection module is used for adopting the real-time posture detection model that passes the test to perform real-time posture detection on small targets in a dense crowd and output real-time posture detection results.

[0035] The modified Alphapose framework is specifically as follows:

[0036] The YOLOV10 algorithm is used to replace the end-side target detection algorithm in the Alphapose framework to obtain a human body detection frame in an image; the human body detection confidence threshold in the YOLOV10 algorithm is set to a minimum value; a non-maximum suppression algorithm based on a CIOU loss function is used for post-processing of the YOLOV10 algorithm to screen n human body detection frames pointing to the same target; the n human body detection frames pointing to the same target are subjected to skeleton point detection by a Fastpose skeleton point detection network to extract n skeleton point detection results of the target; 26 skeleton points corresponding to the n skeleton point detection results are traversed respectively to take skeleton points with maximum confidence to recombine posture detection results corresponding to 26 human body skeleton points.

[0037] The application provides a computer readable storage medium comprising a stored computer program, wherein the computer readable storage medium controls a device in which the computer readable storage medium is located to execute the method for real-time posture detection of small targets in a dense crowd when the computer program is running.

[0038] The application has the following beneficial effects:

[0039] (1) By using a two-stage posture detection Alphapose network, the detector is replaced by YOLOV10n or YOLOV10s, and the detection confidence threshold is set to 0 to retain more small target frames, and finally the NMS of the CIOU loss function is used to filter the repeated frames, so that the long time-consuming problem of inference is solved, the small target loss problem is solved, and the recall rate of the detector is improved.

[0040] (2) Through the combination of multiple Fastpose detection skeleton point results, the skeleton point set obtained by Fastpose detection on multiple incomplete human body detection boxes is extracted, and the highest confidence skeleton point in each point set is extracted to combine into a new human skeleton, which improves the accuracy of the two-stage pose detection on incomplete human body detection and improves the human skeleton point recognition rate of the pose detection algorithm. BRIEF DESCRIPTION OF DRAWINGS

[0041] The application will be further described below with reference to the accompanying drawings.

[0042] Figure 1 The method flowchart of the first embodiment of the application is shown in the figure.

[0043] Figure 2 The comparison chart of the results before and after the screening of the non-maximum suppression algorithm based on the CIOU loss function based on different inputs in the second embodiment of the application is shown in the figure.

[0044] Figure 3 The skeleton point fusion process diagram of the Fastpose pose detection network in the second embodiment of the application is shown in the figure. DETAILED DESCRIPTION

[0045] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0046] Reference Figure 1 As the first embodiment of the application, a small target real-time pose detection method in a dense crowd is disclosed, which includes the following steps:

[0047] S1, establishment: a real-time pose detection model is constructed by using a modified Alphapose framework;

[0048] S2, training: multiple images labeled with poses are divided into a training set, a validation set and a test set;

[0049] S3, testing: the training set and the validation set are used to train the real-time pose detection model until the real-time pose detection model passes the test of the test set;

[0050] S4, detection: the real-time pose detection model that passes the test is used to perform real-time pose detection on small targets in a dense crowd, and output real-time pose detection results;

[0051] The modified Alphapose framework is specifically:

[0052] S11. Use the YOLO V10 algorithm to replace the on-device object detection algorithm in the Alphapose framework to obtain the human detection frame in the image.

[0053] S12. Set the human detection confidence threshold in the YOLOV10 algorithm to the minimum value (i.e., 0);

[0054] S13. Use the non-maximum suppression algorithm (NMS algorithm) based on the CIOU loss function to post-process the YOLOV10 algorithm and screen n human detection boxes pointing to the same target;

[0055] S14, using the Fastpose skeleton point detection network, perform skeleton point detection on n human detection frames pointing to the same target, and extract n skeleton point detection results of the target;

[0056] S15. Traverse the skeleton points corresponding to the 26 points in the n skeleton point detection results, take the skeleton point with the largest confidence, and reassemble it into a posture detection result corresponding to the 26 human skeleton points.

[0057] This embodiment also discloses a device for real-time posture detection of small targets in dense crowds, comprising: a building module, a training module, a testing module, and a detection module;

[0058] The building module is used to build a real-time posture detection model using the modified Alphapose framework;

[0059] The training module is used to use multiple posture-labeled images, which are divided into a training set, a validation set and a test set;

[0060] The testing module is used to train the real-time posture detection model using the training set and the validation set until the real-time posture detection model passes the test of the test set;

[0061] The detection module is used to use a tested real-time posture detection model to perform real-time posture detection on small targets in a dense crowd and output real-time posture detection results;

[0062] This embodiment uses the two-stage pose detection Alphapose network, replaces the detector with YOLOV10n or YOLOV10s, and retains more small target frames by setting the detection confidence threshold to 0. Finally, the NMS with CIOU loss function is used to filter duplicate frames. This not only solves the problems of long inference time and high computing resource consumption, but also solves the problem of small target loss, improves the recall rate of the detector, and provides the possibility for real-time inference of two-stage pose detection embedded terminal deployment.

[0063] By merging the skeleton point results of multiple Fastpose detections, Fastpose detection is performed on multiple incomplete human detection frames to obtain a set of skeleton points. In the set of 26 skeleton points, the skeleton point with the highest confidence for each point in the set is extracted to merge them into a new human skeleton. This improves the accuracy of posture detection for incomplete human bodies in the two-stage method and also improves the human skeleton point recognition rate of the posture detection algorithm.

[0064] See also Figures 2 to 3 As a second embodiment of the present invention, a more specific embodiment of the first embodiment, based on Alphapose, the detector is replaced with YOLOV10n or YOLOV10s, and pose detection uses Fastpose's 26-body skeleton detection network. By reducing the YOLOV10 human detection confidence threshold, more small objects are retained. Then, duplicate boxes are filtered based on the CIOU non-maximum threshold (NMS). Finally, skeleton points pointing to the same object are selected and merged.

[0065] This embodiment improves the Alphapose framework in two steps:

[0066] Step 1: Improvement of detector;

[0067] (11) Replace the detector with YOLOV10n or YOLOV10s.

[0068] YOLOv10 is an object detection algorithm trained using Neural Network Detection (NMS) without post-processing. Use YOLOv10n or YOLOv10s to replace Alphapose's YOLOX-x or YOLOv3-SPP detection algorithms. Compared to YOLOX-x or YOLOv3-SPP, YOLOv10n or YOLOv10s have smaller networks and can achieve several times faster model inference speed.

[0069] (12) The confidence threshold of target detection is minimized to retain more small target boxes.

[0070] After actual testing and verification, when the network and input size are fixed, the larger the confidence threshold, the more serious the loss of small targets. In particular, the confidence of small targets is mostly less than 0.25. When the confidence threshold is fixed, YOLOV10N (640*640) is compared with YOLOV10N (1080*1080), and YOLOV10S (640*640) is compared with YOLOV10S (1080*1080). After actual testing, it is concluded that when recognizing high-resolution network input, although the neural network obtains more detection information with large-size input, some small targets will also be lost when the confidence threshold is less than 0.25. Combined with other actual test results, the confidence threshold is inversely proportional to the loss of small targets in human detection; when the confidence threshold is set to 0, human detection loses the least small targets. Since the confidence threshold is reduced, more duplicate frames will appear when pointing to a target; in order to solve this problem, the present invention uses a non-maximum suppression algorithm (NMS) based on CIOU to filter out the redundant duplicate frames that appear in step (12). The effect after filtering is as follows: Figure 2 shown.

[0071] (13) The human body detection frame performs a similar non-maximum suppression algorithm to filter out frames pointing to the same target.

[0072] Specifically, the sub-step S31 of this embodiment includes the following sub-steps:

[0073] Step a1: export all human detection boxes pointing to the same target in the human detection boxes identified by the YOLOV10 algorithm with the human detection confidence threshold set to the minimum value as a list BoxLists, and create a new list ResultBoxLists for saving the final results;

[0074] Step a2: sort each human detection box in the list BoxLists in descending order of confidence to obtain the list SortBoxLists;

[0075] Step a3: traverse each human detection frame in the list SortBoxLists, calculate the CIOU loss function for the first human detection frame ranked first in the list SortBoxLists and the remaining human detection frames, and record the human detection frames whose CIOU loss function calculation results are greater than 0.25;

[0076] Step a4: insert the human body detection box ranked first in the SortBoxLists list into the ResultBoxLists list;

[0077] Step a5, deleting the first human detection frame in the list SortBoxLists and the human detection frame recorded in step a3 from the list SortBoxLists;

[0078] Step a6, looping through steps a3 to a5 until the number of human body detection boxes in the list SortBoxLists is less than or equal to one;

[0079] Step a7: insert the remaining human detection frames in the list SortBoxLists into the list ResultBoxLists, and output n human detection frames pointing to the target; where n is the number of human detection frames finally included in the list ResultBoxLists.

[0080] In this embodiment, in the non-maximum suppression algorithm based on the CIOU loss function, the calculation formula of the CIOU loss function is as follows:

[0081]

[0082] Where w' and h' are the width and height of the intersection of the two human detection frames, w1, h1, w2, and h2 are the width and height of the two human detection frames, respectively. ρ is the distance between the center points of the two human detection frames, and c is the length of the hypotenuse of the bounding box composed of the two targets (i.e., the leftmost boundary, rightmost boundary, topmost boundary, and bottommost boundary of the two frames are used to synthesize a new frame).

[0083] YOLOV10 is a target detection algorithm trained with NMS without post-processing, achieving better robustness for small targets. When the confidence threshold is set to 0, CIOU NMS filtering is performed, and the inference process is basically the same as the YOLOX-x and YOLOV3-SPP detection algorithms. This embodiment introduces YOLOV10 lightweight network inference without changing the overall process of Alphapose inference, which reduces the model inference time while retaining more small targets.

[0084] Step 2: Posture Detection Improvement

[0085] The posture detection network used in the present invention is Fastpose. The detection results of step 1 are obtained, multiple target frames pointing to the same person are screened out, posture detection is performed, and then the posture skeleton points are merged.

[0086] (21) 26 Fastpose skeleton point detection algorithms.

[0087] Fastpose is an algorithm that detects 26 human skeleton points. The 26 skeleton points cover the entire human body from the top of the head to the feet.

[0088] (22) Using CIOU’s NMS algorithm, we filter boxes pointing to the same target.

[0089] (23) Fastpose detection is performed on a set of boxes pointing to the same target, and then skeleton points are fused.

[0090] Traverse the list PointsLists of boxes pointing to the same target in the above step (21), perform Fasepose skeleton point detection on the n target boxes pointing to the same person, and obtain n*26 skeleton points.

[0091] If n is 1, the 26 skeleton points obtained are the final skeleton points of the human body. If n is greater than 1, the 26 skeleton points are traversed, and the point with the highest confidence among the n 26 skeleton points is taken as the point of the human body at the corresponding position, and reassembled into 26 human skeleton points.

[0092] In other embodiments, sub-step S31 may further include the following sub-steps:

[0093] The IOU loss function is calculated for all human detection frames that point to the same target in the human detection frame identified by the YOLOV10 algorithm with the human detection confidence threshold set to the minimum value. The human detection frames with a calculated result greater than 0.8 are merged into a new human detection frame; that is, the leftmost boundary, rightmost boundary, topmost boundary, and bottommost boundary of the two frames are taken to synthesize the new human detection frame.

[0094] That is, the CIOU NMS filtering algorithm for the human detection frame when the confidence threshold is 0 can be replaced by merging the human detection frame with a confidence threshold of 0 with overlapping (the IOU of two frames is greater than 0.8) (that is, taking the leftmost boundary, rightmost boundary, topmost boundary and bottommost boundary of two human detection frames to synthesize a new human detection frame).

[0095] The present invention also discloses a terminal device, including a processor and a storage device, wherein the storage device is used to store one or more programs; when the one or more programs are executed by the processor, the processor implements the above-mentioned method for real-time posture detection of small targets in dense crowds. The processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc. The processor is the control center of the test equipment, and uses various interfaces and lines to connect the various parts of the entire test equipment.

[0096] The storage device can be used to store computer programs and / or modules. The processor implements various functions of the terminal device by running or executing the computer programs and / or modules stored in the storage device and calling data stored in the storage device. The storage device may mainly include a program storage area and a data storage area. The program storage area may store an operating system, at least one application required for a function, etc.; the data storage area may store data created based on the use of the terminal device, etc. In addition, the storage device may include high-speed random access memory and non-volatile memory, such as a hard disk, internal memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0097] The device integration module / unit for real-time small target posture detection in dense crowds can be stored in a computer-readable storage medium if it is implemented in the form of a software function unit and sold or used as an independent product. Based on this understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in at least one computer-readable storage medium. When the processor executes the computer program, the steps of the above-mentioned various method embodiments can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, an executable file, or some intermediate form. The computer-readable medium can include any entity or device capable of carrying computer program code, a recording medium, a U disk, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0098] It should be noted that the above-described embodiments of the device and apparatus are only illustrative, and the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e., they can be located in one place or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the present embodiment scheme.

Claims

1. A method for real-time posture detection of small targets in dense crowds, characterized by: include: A real-time posture detection model is constructed using the modified Alphapose framework; Use multiple pose-labeled images and divide them into training set, validation set and test set; The real-time posture detection model is trained using the training set and the validation set until the real-time posture detection model passes the test of the test set; Use the tested real-time posture detection model to perform real-time posture detection on small targets in dense crowds and output real-time posture detection results; The modified Alphapose framework is specifically: Use the YOLO V10 algorithm to replace the on-device object detection algorithm in the Alphapose framework to obtain human detection frames in the image. Set the human detection confidence threshold in the YOLOV10 algorithm to the minimum value; The non-maximum suppression algorithm based on the CIOU loss function is used for post-processing of the YOLOV10 algorithm to screen n human detection boxes pointing to the same target; Through the Fastpose skeleton point detection network, skeleton point detection is performed on n human detection boxes pointing to the same target, and n skeleton point detection results of the target are extracted; Traverse the 26 points in the n skeleton point detection results that correspond to the skeleton points, take the skeleton point with the highest confidence, and reassemble it into the posture detection result corresponding to the 26 human skeleton points; The non-maximum suppression algorithm based on the CIOU loss function is used to post-process the YOLOV10 algorithm to screen n human detection frames pointing to the same target. The specific steps are as follows: Step a1: export all human detection boxes pointing to the same target in the human detection boxes identified by the YOLOV10 algorithm with the human detection confidence threshold set to the minimum value as a list BoxLists, and create a new list ResultBoxLists for saving the final results; Step a2: sort each human detection box in the list BoxLists in descending order of confidence to obtain the list SortBoxLists; Step a3: traverse each human detection frame in the list SortBoxLists, calculate the CIOU loss function for the first human detection frame ranked first in the list SortBoxLists and the remaining human detection frames, and record the human detection frames whose CIOU loss function calculation results are greater than 0.25; Step a4: insert the human body detection box ranked first in the SortBoxLists list into the ResultBoxLists list; Step a5, deleting the first human detection frame in the list SortBoxLists and the human detection frame recorded in step a3 from the list SortBoxLists; Step a6, looping through steps a3 to a5 until the number of human body detection boxes in the list SortBoxLists is less than or equal to one; Step a7: insert the remaining human detection frames in the list SortBoxLists into the list ResultBoxLists, and output n human detection frames pointing to the target; where n is the number of human detection frames finally included in the list ResultBoxLists; In the non-maximum suppression algorithm based on the CIOU loss function, the calculation formula of the CIOU loss function is as follows: Where w' and h' are the width and height of the intersection of the two human detection frames, w1, h1, w2 and h2 are the width and height of the two human detection frames respectively, ρ is the distance between the center points of the two human detection frames, and c is the length of the hypotenuse of the bounding box composed of the two targets; The method uses the non-maximum suppression algorithm based on the CIOU loss function to post-process the YOLOV10 algorithm to screen n human detection frames pointing to the same target, and also includes the following steps: In the human detection frame identified by the YOLOV10 algorithm with the human detection confidence threshold set to the minimum value, all human detection frames pointing to the same target are calculated, and the IOU loss function is calculated. The human detection frames with a calculation result greater than 0.8 are merged into a new human detection frame.

2. The method for real-time posture detection of small targets in a dense crowd according to claim 1, characterized in that: The YOLOV10 algorithm specifically adopts one of the YOLOV10n algorithm or the YOLOV10s algorithm.

3. A device for real-time posture detection of small targets in dense crowds, characterized by: include: Building modules, training modules, testing modules and detection modules; The building module is used to build a real-time posture detection model using the modified Alphapose framework; The training module is used to use multiple posture-labeled images, which are divided into a training set, a validation set and a test set; The testing module is used to train the real-time posture detection model using the training set and the validation set until the real-time posture detection model passes the test of the test set; The detection module is used to use a tested real-time posture detection model to perform real-time posture detection on small targets in a dense crowd and output real-time posture detection results; The modified Alphapose framework is specifically: The YOLO V10 algorithm is used to replace the end-to-end object detection algorithm in the Alphapose framework to obtain human detection frames in the image. The human detection confidence threshold in the YOLO V10 algorithm is set to the minimum value. The non-maximum suppression algorithm based on the CIOU loss function is used for post-processing of the YOLO V10 algorithm to screen n human detection frames pointing to the same target. The Fastpose skeleton point detection network is used to perform skeleton point detection on each of the n human detection frames pointing to the same target, and n skeleton point detection results for the target are extracted. The skeleton points corresponding to the 26 points in the n skeleton point detection results are traversed, and the skeleton point with the highest confidence is selected, which is recombined into a posture detection result corresponding to the 26 human skeleton points. The non-maximum suppression algorithm based on the CIOU loss function is used to post-process the YOLOV10 algorithm to screen n human detection frames pointing to the same target. The specific steps are as follows: Step a1: export all human detection boxes pointing to the same target in the human detection boxes identified by the YOLOV10 algorithm with the human detection confidence threshold set to the minimum value as a list BoxLists, and create a new list ResultBoxLists for saving the final results; Step a2: sort each human detection box in the list BoxLists in descending order of confidence to obtain the list SortBoxLists; Step a3: traverse each human detection frame in the list SortBoxLists, calculate the CIOU loss function for the first human detection frame ranked first in the list SortBoxLists and the remaining human detection frames, and record the human detection frames whose CIOU loss function calculation results are greater than 0.25; Step a4: insert the human body detection box ranked first in the SortBoxLists list into the ResultBoxLists list; Step a5, deleting the first human detection frame in the list SortBoxLists and the human detection frame recorded in step a3 from the list SortBoxLists; Step a6, looping through steps a3 to a5 until the number of human body detection boxes in the list SortBoxLists is less than or equal to one; Step a7: insert the remaining human detection frames in the list SortBoxLists into the list ResultBoxLists, and output n human detection frames pointing to the target; where n is the number of human detection frames finally included in the list ResultBoxLists; In the non-maximum suppression algorithm based on the CIOU loss function, the calculation formula of the CIOU loss function is as follows: Where w' and h' are the width and height of the intersection of the two human detection frames, w1, h1, w2 and h2 are the width and height of the two human detection frames respectively, ρ is the distance between the center points of the two human detection frames, and c is the length of the hypotenuse of the bounding box composed of the two targets; The method uses the non-maximum suppression algorithm based on the CIOU loss function to post-process the YOLOV10 algorithm to screen n human detection frames pointing to the same target, and also includes the following steps: In the human detection frame identified by the YOLOV10 algorithm with the human detection confidence threshold set to the minimum value, all human detection frames pointing to the same target are calculated, and the IOU loss function is calculated. The human detection frames with a calculation result greater than 0.8 are merged into a new human detection frame.

4. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the method for real-time posture detection of small targets in a dense crowd as described in any one of claims 1 to 2.

Images