Prediction method of badminton landing position based on computer vision

Abstract

The invention relates to a badminton landing position prediction method based on computer vision. According to a video frame picture obtained from a badminton match video, the perspective transformation relationship between a badminton field and a model field in the video frame picture is calculated according to the video frame picture obtained from the badminton match video, and a badminton fieldmodel is determined. Then the motion region detection module uses the time difference method to distinguish the motion region and the background part. Binary image processing, image segmentation, contour tracking are performed on a difference image and then the characteristic information of badminton is extracted; the flight trajectory of badminton is determined by trajectory tracking algorithm;Kalman filter algorithm is used to predict the trajectory location of badminton. According to the location of trajectory falling point, the coordinate of actual badminton court is transformed by perspective transformation, so as to realize the prediction of the location of badminton falling point. The method can be used to predict the badminton landing position in the complex scene when the illumination changes gradually, the wind speed influence, the airflow influence or the badminton speed is very fast.

Description

technical field [0001] The invention belongs to the technical fields of digital image processing, pattern recognition and machine learning, and video target detection, and in particular relates to a method for predicting the position of a badminton drop point by using computer vision technology. Background technique [0002] The method of using computer vision technology to predict the location of badminton landing can be applied to sports games and Olympic badminton competitions. The main function is that the system can quickly identify the badminton competition venue and the flying shuttlecock according to the badminton flight trajectory of the input badminton competition video. , to predict where the shuttlecock will land. In the method of predicting the position of the badminton landing point, a series of operations such as image preprocessing, image recognition, image segmentation, moving object detection, object tracking, and motion trajectory prediction are performed ...

AI Technical Summary

Problems solved by technology

Therefore, in recent years, many scholars have carried out in-depth research on target tracking and proposed many algorithms; however, the tracking of the specific target of badminton is easily affected by factors such as light and wind speed, making the prediction of the location of the badminton landing still difficult. is a challenge

Images