Research On Key Technology Of Ball Picking Robot Based On Monocular Vision

The development and progress of science and technology is aimed at improving the quality of human life.The industrialization and achievement of advanced science and technology have far-reaching influence and significance for human production and life.The research and development of mobile robot technology is of great significance to the improvement of human life quality.This paper takes the croquet robot based on monocular vision as the research object,and studies its key technologies.The main work is as follows:(1)Aiming at the problem of the recognition of the spherical target by the croquet robot,the final recognition of the spherical target is completed by matching the two feature quantities of the color and shape of the spherical target.In the process of recognizing the color features of the sphere,the color space model is transformed,and the HSV color space model adopts the method of double threshold for color recognition.In the process of identifying the shape feature quantity,after grayscale processing of color images,the problem of inaccurate image segmentation under the condition of uneven illumination is improved by adaptive threshold segmentation,after the Candy edge detection method is used to extract the edges,the shape is recognized by hough circle transformation.Finally,experiments show that this method is ideal for the recognition of target sphere under different illumination conditions.(2)In-depth study of Zhang calibration method adopted in this paper,in order to ensure the accuracy of the calibration results,images of different postures of multiple calibration plates are collected,finally,the internal parameter matrix and distortion coefficient of the camera are obtained.bCombined with the actual working background of the Ryukyu robot,the near-field constraint model is used to perform the monocular vision ranging experiment on the target sphere.Through the analysis of experimental results,the ranging algorithm satisfies the ranging requirement of the target sphere within a certain range.(3)Based on the limited field of vision of ball-picking robot,a hierarchical planning idea is used to divide the global window,in each window,ant colony algorithm is adopted to carry out multi-objective path planning for the balls in the Windows,Matlab simulation results show that although the global optimization cannot be achieved after the robot search is complete,it is close to the global optimization and has strong applicability to the characteristics of the ball-picking robot.Aiming at the obstacle avoidance problem in the working environment of ball picking robot,based on the traditional artificial potential field method,the relative velocity between robot and obstacle and the distance adjustment factor is introduced,the angle of potential field force is adjusted,add the adaptive step size,Matlab simulation results show that this algorithm can sove the problem of inaccessible target point of robot under special position relationship and the problem of obstacle avoidance trajectory is not smooth,and robots can get rid of obstacles faster.