Return trajectory prediction method based on binocular vision perception of racket swinging and ball hitting actions

Abstract

The invention discloses a return trajectory prediction method based on binocular vision perception of racket swinging and ball hitting actions. According to the method, the track of the racket beforebatting is used as an important basis for predicting the track of the shuttlecock, and the predicted time point is lifted from the time after the racket leaves the shuttlecock to the time before the racket touches the shuttlecock, so that more time is gained, and the quick response capability, batting point optimization capability, batting initiative and competitive level of the shuttlecock robotare improved. The method specifically comprises the following steps: 1, processing; first, a binocular camera is calibrated, denoising and distortion correction are carried out on the collected batting binocular image; identifying a motion area through a frame difference method; the shuttlecock and the racket are detected by using contour shape and area characteristics; eXTRACTING CORRESPONDING FEATURE POINTS, the method comprises the steps of matching binocular images according to an epipolar constraint principle, calculating parallax and depth of feature points, solving coordinates of the feature points, screening out off-site invalid moving targets according to boundary lines, calculating effective target motion parameters, taking the effective target motion parameters as neural networkinput for learning or prediction, and finally performing prediction calculation on medium and long term trajectories of return balls by utilizing Kalman filtering.

Description

technical field [0001] The invention belongs to the field of image information technology, specifically, in badminton, based on the binocular image sequence analysis of the swinging action, the method for predicting the trajectory of the incoming ball can be used for the visual perception of the badminton robot system or the digital auxiliary training of badminton . Background technique [0002] In recent years, vision-based target detection, tracking and trajectory prediction methods have emerged rapidly and been widely used. The prediction of badminton trajectory can be used for auxiliary training of players, and can also be used as necessary information input for badminton robots. The patent No. CN109087327A discloses a computer vision-based method for predicting the position of the shuttlecock's landing point. [0003] Badminton is an extremely fast game against the net with rackets. In competition, it is not only necessary to judge the trajectory of the incoming ball...

AI Technical Summary

Problems solved by technology

However, when the ball speed is fast or the flight distance is short, the calculation, communication, and execution systems often fail to respond, resulting in failure to hit the ball

Images