Artificial Intelligence-Based Method and System for Detecting Violations and Analyzing Movements in Shot Put

Abstract

This application discloses a method and system for detecting fouls and analyzing actions of shot put throwing based on artificial intelligence. First, a video containing multiple frames of video images of the person under test throwing the shot put collected by a camera is obtained; for each frame of video image in the video, the human body region of the person under test and the bounding box of the shot put are obtained through an object detection algorithm; then, the first current frame image in which the distance between the human body region and the bounding box is just greater than a preset threshold is determined, and the second current frame image in which the human body region and the bounding box are just separated is determined; finally, foul detection is performed based on the previous frame image of the first current frame image, and throwing analysis is performed on the person under test based on the second current frame image and its previous frame image. The present invention uses artificial intelligence and machine vision technologies to achieve foul detection and action analysis under the conditions of no supervision and no guidance, reducing labor consumption and improving test efficiency.

Description

Technical Field

[0001] The present invention relates to the field of motion detection and analysis, and in particular to a shot put violation detection and motion analysis method and system based on artificial intelligence. Background Art

[0002] Shot put is a track and field throwing event that tests strength and technique. When an athlete throws the shot put, they must follow the rules and the shot put must be thrown from the shoulder with one hand and the athlete's feet must be within the throwing circle.

[0003] In the prior art, athletes are usually judged and their movements analyzed manually. However, manual judgment and analysis are greatly affected by environmental factors and are prone to misjudgment. Summary of the invention

[0004] Based on this, the embodiment of the present application provides an artificial intelligence-based shot put violation detection and motion analysis method and system, which can realize violation detection and motion analysis under unsupervised and unguided conditions, reduce manpower consumption, improve testing efficiency, and improve the accuracy of judgment analysis.

[0005] In a first aspect, a shot put violation detection and motion analysis method based on artificial intelligence is provided, the method comprising:

[0006] Obtain a video containing multiple frames of video images of the test person throwing the shot put captured by the camera;

[0007] For each frame of video image in the video, a human body region of the tested person and a bounding box of the shot put are obtained by using a target detection algorithm;

[0008] Determine a first current frame image in which the distance between the human body region and the bounding box is just greater than a preset threshold; wherein the distance between the human body region and the bounding box in a frame image previous to the first current frame image is less than or equal to the preset threshold;

[0009] Determine a second current frame image in which the human body region and the boundary frame are just separated; wherein the human body region and the boundary frame in a previous frame image of the second current frame image have an overlapping area;

[0010] Performing violation detection on the tested person based on the previous frame image of the first current frame image to obtain a violation detection result;

[0011] Based on the second current frame image and the previous frame image, a pushing and throwing analysis is performed on the subject, and a pushing and throwing suggestion is provided based on the pushing and throwing analysis.

[0012] Optionally, performing violation detection on the tested person based on a previous frame image of the first current frame image includes:

[0013] For the previous frame image of the first current frame image, obtain the key point information of the human body through the human body pose key point detection algorithm, and obtain the bounding box of the shot put according to the object detection algorithm;

[0014] Judge whether the tested person has a violation of pushing the shot put under the shoulder or pushing it backward according to the key point information and the bounding box;

[0015] And identify whether there is an intersection between the boundary of the tested person's foot and the throwing circle through the object detection algorithm, and judge whether the tested person has a foul action of touching the throwing circle.

[0016] Optionally, perform throwing analysis on the tested person based on the second current frame image and its previous frame image, including:

[0017] Establish a plane coordinate system based on the second current frame image and its previous frame image;

[0018] According to the time interval between two frames and the center coordinates of the shot put in the two frames, calculate the speed at which the shot put is thrown, and determine the throwing direction according to the angle between the line connecting the center coordinates of the shot put in the two frames and the ground.

[0019] Optionally, provide throwing suggestions according to the speed at which the shot put is thrown and the throwing direction.

[0020] Optionally, obtain the video image data of the target person throwing the shot put collected by the camera, including:

[0021] Collect image data video image data at 25 frames per second with a high-definition camera, that is, the time interval from the previous frame to the next frame is 4 ms.

[0022] In a second aspect, a shot put throwing violation detection and action analysis system based on artificial intelligence is provided, and the system includes:

[0023] A video acquisition module for acquiring a video containing multiple frame video images of the tested person throwing the shot put collected by the camera;

[0024] An object detection module for, for each frame of video image in the video, obtaining the human body area of the tested person and the bounding box of the shot put through the object detection algorithm;

[0025] A key frame extraction module, configured to determine a first current frame image in which the distance between the human body region and the bounding box is just greater than a preset threshold; wherein, in the previous frame image of the first current frame image, the distance between the human body region and the bounding box is less than or equal to the preset threshold; determine a second current frame image in which the human body region and the bounding box are just separated; wherein, in the previous frame image of the second current frame image, there is an overlapping region between the human body region and the bounding box.

[0026] A violation detection module, configured to perform violation detection on the person being tested based on the previous frame image of the first current frame image, and obtain a violation detection result.

[0027] An action analysis module, configured to perform a throwing analysis on the person being tested based on the second current frame image and its previous frame image, and provide a throwing suggestion according to the throwing analysis.

[0028] Optionally, the violation detection module includes:

[0029] For the previous frame image of the first current frame image, obtain the key point information of the human body through a human pose key point detection algorithm, and obtain the bounding box of the shot put according to the object detection algorithm.

[0030] Judge whether the person being tested has a violation action of throwing the shot put under the shoulder or throwing it backward according to the key point information and the bounding box.

[0031] And identify whether there is an intersection between the boundary of the feet of the person being tested and the throwing circle through the object detection algorithm, and judge whether the person being tested has a foul action of touching the throwing circle.

[0032] Optionally, the action analysis module includes:

[0033] Establish a plane coordinate system based on the second current frame image and its previous frame image.

[0034] According to the time interval between two frames and the center coordinates of the shot put in the two frames, calculate the speed at which the shot put is thrown, and determine the throwing direction according to the angle between the line connecting the center coordinates of the shot put in the two frames and the ground.

[0035] Optionally, provide a throwing suggestion according to the speed at which the shot put is thrown and the throwing direction.

[0036] Optionally, the video acquisition module includes:

[0037] Collect image data video images at 25 frames per second using a high-definition camera, that is, the time interval from the previous frame to the next frame is 4 ms.

[0038] In the technical solution provided by the embodiment of the present application, first, a video including multiple video images of a tested person throwing a shot put collected by a camera is obtained; for each video image in the video, the human body area of the tested person and the bounding box of the shot put are obtained through a target detection algorithm; the first current frame image in which the distance between the human body area and the bounding box is just greater than a preset threshold is determined; the second current frame image in which the human body area and the bounding box are just separated is determined; the illegal behavior of the tested person is detected based on the previous frame image of the first current frame image to obtain an illegal detection result; the throwing of the tested person is analyzed based on the second current frame image and its previous frame image, and a throwing suggestion is provided according to the throwing analysis. It can be seen that the beneficial effect of the present invention is that by using artificial intelligence and machine vision technologies, illegal behavior detection and motion analysis can be realized under the conditions of no supervision and no guidance, reducing labor consumption and improving the test efficiency. Description of the Drawings

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only exemplary, and for those of ordinary skill in the art, without creative efforts, other implementation drawings can be obtained according to the provided drawings.

[0040] Figure 1 It is a flowchart of a shot put throwing illegal behavior detection and motion analysis based on artificial intelligence provided by the embodiment of the present application;

[0041] Figure 2 It is a block diagram of a shot put throwing illegal behavior detection and motion analysis system based on artificial intelligence provided by the embodiment of the present application. Detailed Embodiments

[0042] In order to make the purpose, technical solutions and advantages of the present application more clear, the following further details the present application in combination with the drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application, and are not used to limit the present application.

[0043] In the description of the present invention, unless otherwise specified, the meaning of "a plurality" is two or more. The terms "first", "second", etc. in the specification and claims of the present invention and the above drawings are intended to distinguish the objects being referred to. For a solution with a time sequence process, this way of term expression does not necessarily need to be understood as describing a specific order or sequence, and for a solution of a device structure, this way of term expression also does not distinguish the importance level, positional relationship, etc.

[0044] In addition, the terms "comprising", "having", and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units need not be limited to those steps or units that are explicitly listed, but may also include other steps or units that are inherent to these processes, methods, products, or devices although not explicitly listed, or steps or units added based on further optimization schemes of the inventive concept of the present invention.

[0045] Please refer to Figure 1 , which shows a flowchart of a method for detecting fouls and analyzing actions of shot put throwing based on artificial intelligence provided by an embodiment of the present application. The method may include the following steps:

[0046] Step 101, obtain a video including multiple frames of video images of a person under test throwing a shot put collected by a camera.

[0047] In an embodiment of the present application, high-definition camera is used to collect image data of video images at 25 frames per second, that is, the time interval from the previous frame to the next frame is 4 ms.

[0048] Step 102, for each frame of video image in the video, obtain the human body area of the person under test and the bounding box of the shot put through a target detection algorithm;

[0049] Step 103, determine the first current frame image in which the distance between the human body area and the bounding box is just greater than a preset threshold; determine the second current frame image in which the human body area and the bounding box are just separated;

[0050] Wherein, the distance between the human body area and the bounding box in the previous frame image of the first current frame image is less than or equal to the preset threshold; there is an overlapping area between the human body area and the bounding box in the previous frame image of the second current frame image.

[0051] Specifically, each frame of the video image data of the person under test throwing a shot put obtained by the video acquisition device is analyzed, the human body information of the person under test and the bounding box of the shot put are obtained through the target detection algorithm, and the current frame, the previous frame when the distance between the human body and the bounding box is greater than a specific threshold a, and the previous frame and the current frame when the shot put leaves the body are found, and these two frames of images are sent to the foul detection module and the action analysis module.

[0052] Step 104, perform foul detection on the person under test based on the previous frame image of the first current frame image to obtain a foul detection result; perform shot put analysis on the person under test based on the second current frame image and its previous frame image, and provide a shot put throwing suggestion according to the shot put analysis.

[0053] In the embodiment of the present application, the illegal detection of the person under test is performed based on the previous frame image of the first current frame image, including obtaining the key point information of the human body through the human body pose key point detection algorithm in the previous frame image of the first current frame image, and the bounding box of the shot put obtained according to the object detection algorithm; judging whether the person under test has illegal actions such as pushing the shot put under the shoulder or pushing it backward according to the key point information and the bounding box; and identifying whether there is an intersection between the boundary of the foot of the person under test and the throwing circle through the object detection algorithm, and judging whether the person under test has a foul action of touching the throwing circle.

[0054] Specifically, according to the previous frame extracted by the key frame extraction module, the key point information of the human body is obtained through the human body pose key point detection algorithm, and the bounding box of the shot put obtained by the object detection algorithm. According to the position information of the key points and the bounding box on the image, it is judged whether there are illegal actions such as pushing the shot put under the shoulder or pushing it backward, and whether there is an intersection between the bounding box of the foot of the person under test and the throwing circle is identified through the object detection algorithm, and it is judged whether any part of the body touches the throwing circle for a foul action.

[0055] The throwing analysis of the person under test is performed based on the second current frame image and its previous frame image, including establishing a plane coordinate system based on the second current frame image and its previous frame image; calculating the speed at which the shot put is thrown out according to the time interval between the two frames and the center coordinates of the shot put in the two frames, and determining the throwing direction according to the angle between the line connecting the center coordinates of the shot put in the two frames and the ground. Throwing suggestions are provided according to the speed at which the shot put is thrown out and the throwing direction.

[0056] Specifically, according to the previous frame and the current frame extracted by the key frame extraction module, the object detection algorithm obtains the bounding boxes of the shot put in these two frames, and takes a vertex of the image as the origin, and the edge lines extending horizontally and vertically as the x-axis and y-axis, so as to establish a plane coordinate system, and calculate the position of the center point of the bounding box (xi, yi) i = 1, 2 accordingly. And calculate the speed at which the shot put is thrown out according to the time interval of 4 ms between the two frames and the distance from (x1, y1) to (x2, y2), and take the angle between the line connecting (x1, y1) to (x2, y2) and the x-axis as the throwing direction, and provide throwing suggestions according to the throwing speed and direction.

[0057] Such as Figure 2 , the embodiment of the present application also provides an artificial intelligence-based shot put illegal detection and action analysis system 200. The system 200 includes:

[0058] A video acquisition module 201, configured to acquire a video including multiple frame video images of the person under test throwing the shot put collected by the camera;

[0059] The target detection module 202 is configured to obtain the human body region of the person to be tested and the bounding box of the shot put in each frame of video image in the video through a target detection algorithm;

[0060] The key frame extraction module 203 is configured to determine a first current frame image in which the distance between the human body region and the bounding box is just greater than a preset threshold; wherein, the distance between the human body region and the bounding box in the previous frame image of the first current frame image is less than or equal to the preset threshold; determine a second current frame image in which the human body region and the bounding box are just separated; wherein, there is an overlapping region between the human body region and the bounding box in the previous frame image of the second current frame image;

[0061] The violation detection module 204 performs a violation detection on the person to be tested based on the previous frame image of the first current frame image to obtain a violation detection result;

[0062] The action analysis module 205 performs a throwing analysis on the person to be tested based on the second current frame image and its previous frame image, and provides a throwing suggestion according to the throwing analysis.

[0063] In an optional embodiment of the present application, the violation detection module includes, for the previous frame image of the first current frame image, obtaining key point information of the human body through a human body pose key point detection algorithm, and the bounding box of the shot put obtained according to the target detection algorithm; judging whether the person to be tested has a violation action of throwing the shot put under the shoulder or throwing it backward according to the key point information and the bounding box; and identifying whether there is an intersection between the boundary of the foot of the person to be tested and the throwing circle through the target detection algorithm, and judging whether the person to be tested has a foul action of touching the throwing circle.

[0064] In an optional embodiment of the present application, the action analysis module includes establishing a plane coordinate system based on the second current frame image and its previous frame image; calculating the speed at which the shot put is thrown out according to the time interval between the two frames and the center coordinates of the shot put in the two frames, and determining the throwing direction according to the angle between the line connecting the center coordinates of the shot put in the two frames and the ground.

[0065] In an optional embodiment of the present application, a throwing suggestion is provided according to the speed at which the shot put is thrown out and the throwing direction.

[0066] In an optional embodiment of the present application, the video acquisition module includes acquiring video image data of image data at 25 frames per second using a high-definition camera, that is, the time interval from the previous frame to the next frame is 4 ms.

[0067] The system for detecting fouls and analyzing actions in shot put throwing based on artificial intelligence provided by the embodiments of the present application is used to implement the above-mentioned method for detecting fouls and analyzing actions in shot put throwing based on artificial intelligence. For the specific limitations of the system for detecting fouls and analyzing actions in shot put throwing based on artificial intelligence, reference can be made to the limitations of the method for detecting fouls and analyzing actions in shot put throwing based on artificial intelligence in the above text, which will not be elaborated here. Each part of the above-mentioned system for detecting fouls and analyzing actions in shot put throwing based on artificial intelligence can be implemented in whole or in part by software, hardware, and their combinations. The above-mentioned modules can be embedded in the processor of the device in the form of hardware or be independent of it, or be stored in the memory of the device in the form of software, so as to facilitate the processor to call and execute the operations corresponding to the above-mentioned modules.

[0068] The technical features of the above-mentioned embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0069] The above-mentioned embodiments only represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be understood as a limitation on the scope of the patent application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. An illegal throwing detection and motion analysis method for shot put based on artificial intelligence, characterized in that, the method includes: Obtaining a video containing multiple frames of video images of the tested person throwing a shot put collected by a camera; For each frame of video image in the video, obtaining the human body area of the tested person and the bounding box of the shot put through an object detection algorithm; Determining the first current frame image in which the distance between the human body area and the bounding box is just greater than a preset threshold; wherein, the distance between the human body area and the bounding box in the previous frame image of the first current frame image is less than or equal to the preset threshold; Determining the second current frame image in which the human body area and the bounding box are just separated; wherein, there is an overlapping area between the human body area and the bounding box in the previous frame image of the second current frame image; Performing illegal throwing detection on the tested person based on the previous frame image of the first current frame image to obtain an illegal throwing detection result; Performing throwing analysis on the tested person based on the second current frame image and its previous frame image, and providing throwing suggestions according to the throwing analysis; The performing illegal throwing detection on the tested person based on the previous frame image of the first current frame image includes: For the previous frame image of the first current frame image, obtaining the key point information of the human body through a human body pose key point detection algorithm, and the bounding box of the shot put obtained according to the object detection algorithm; Judging whether the tested person has illegal actions such as throwing the shot put under the shoulder or throwing it backward according to the key point information and the bounding box; And identifying whether there is an intersection between the boundary of the tested person's foot and the throwing circle through the object detection algorithm, and judging whether the tested person has a foul action of touching the throwing circle; The performing throwing analysis on the tested person based on the second current frame image and its previous frame image includes: Establishing a plane coordinate system based on the second current frame image and its previous frame image; Calculating the speed at which the shot put is thrown out according to the time interval between two frames and the center coordinates of the shot put in the two frames, and determining the throwing direction according to the angle between the line connecting the center coordinates of the shot put in the two frames and the ground.

2. The method according to claim 1, characterized in that, Providing throwing suggestions according to the speed at which the shot put is thrown out and the throwing direction.

3. The method according to claim 1, characterized in that, The obtaining the video image data of the target person throwing a shot put collected by the camera includes: Collecting image data video image data at 25 frames per second using a high-definition camera, that is, the time interval from the previous frame to the next frame is 4 ms.

4. An illegal throwing detection and motion analysis system for shot put based on artificial intelligence, characterized in that, the system includes: A video acquisition module for obtaining a video containing multiple frames of video images of the tested person throwing a shot put collected by a camera; An object detection module for obtaining the human body area of the tested person and the bounding box of the shot put in each frame of video image in the video through an object detection algorithm; A key frame extraction module, configured to determine a first current frame image in which the distance between the human body region and the bounding box is just greater than a preset threshold; wherein, in the previous frame image of the first current frame image, the distance between the human body region and the bounding box is less than or equal to the preset threshold; determine a second current frame image in which the human body region and the bounding box are just separated; wherein, in the previous frame image of the second current frame image, there is an overlapping region between the human body region and the bounding box. A violation detection module, configured to perform violation detection on the person being tested based on the previous frame image of the first current frame image, and obtain a violation detection result. An action analysis module, configured to perform a throwing analysis on the person being tested based on the second current frame image and its previous frame image, and provide a throwing suggestion according to the throwing analysis. The violation detection module includes: For the previous frame image of the first current frame image, obtain the key point information of the human body through a human pose key point detection algorithm, and obtain the bounding box of the shot put according to the object detection algorithm. Judge whether the person being tested has a violation action of throwing the shot put under the shoulder or throwing it backward according to the key point information and the bounding box. And identify whether there is an intersection between the boundary of the person's foot and the throwing circle through the object detection algorithm, and judge whether the person being tested has a foul action of touching the throwing circle. The action analysis module includes: Establish a plane coordinate system based on the second current frame image and its previous frame image. Calculate the speed at which the shot put is thrown out according to the time interval between the two frames and the center coordinates of the shot put in the two frames, and determine the throwing direction according to the angle between the line connecting the center coordinates of the shot put in the two frames and the ground.

5. The system according to claim 4, wherein, Provide a throwing suggestion according to the speed at which the shot put is thrown out and the throwing direction.

6. The system according to claim 4, wherein, The video acquisition module includes: Collect video image data at 25 frames per second using a high-definition camera, that is, the time interval from the previous frame to the next frame is 4 ms.

Images