Font Size: a A A

Research And Implementation Of Reversing Radar Based On Structured Light

Posted on:2020-06-07Degree:MasterType:Thesis
Country:ChinaCandidate:Y LiFull Text:PDF
GTID:2428330602968361Subject:Circuits and Systems
Abstract/Summary:PDF Full Text Request
At present,reversing assist system extensive used in vehicles has disadvantages.For example,the ultrasonic reversing radar is difficult to sense for low obstacles,the reversing image is distorted,and it is difficult for the driver to grasp the obstacle distance,ground obstacles and wall obstacles cause a great threat to the safety of vehicles during reversing.In this paper,the obstacles in the reversing process are detected by the monocular vision method based on structured light,which can accurately measure the parameters of the obstacles and provide effective information for the driver.Aiming at the shortcomings of the number of convolutions and the large amount of calculation in the process of extracting the center of structured light strips by Hessian matrix method,a method of extracting the center of structured light strips based on gray center of gravity method and Hessian matrix method is proposed.The algorithm extracts the region of interest under the HSV color space of the target image and performs morphological processing,then extracts the initial center of the light strip using the gray center of gravity method,the initial center of the light strip and a certain number of pixels on both sides of its normal direction is calculated as the base point of the second-order Taylor expansion in the Hessian matrix method to obtain an accurate center of the strip.Aiming at the problem that the traditional camera calibration method is difficult to find the global optimal solution and the calibration accuracy is low,a camera calibration method based on improved hybrid particle swarm optimization algorithm is proposed.The method uses Zhang Zhengyou calibration method to calculate the initial value of camera parameters.The initial values of camera parameters are optimized by the combination of adaptive weighting method,simulated annealing algorithm and particle swarm optimization algorithm to obtain the optimal solution of camera parameters.Aiming at the reversing radar based on structured light designed in this paper,the structural light strip center extraction experiment,the reversing radar calibration experiment,the convex obstacle detection experiment and the wall obstacle detection experiment were carried out.The experimental results show that for convex obstacles,the distance detection error is within 69 mm,the width detection error is within 25 mm,and the angle detection error is within 3°;for wall obstacles,the distance detection error is within 45 mm.
Keywords/Search Tags:Reversing radar, Obstacle detection, Structured light strip center extraction, Camera calibration, Improved hybrid particle swarm optimization
PDF Full Text Request
Related items