Three-dimensional Reconstruction And Path Planning Based On Binocular Stereo Vision

As one of the important branches of computer vision,binocular stereo vision is widely used in robot navigation,3D non-contact measurement and virtual reality by simulating the objective world of human vision.Therefore,the research on the three-dimensional reconstruction based on binocular stereo vision and its application in robot path planning has certain theoretical value and very important practical significance.This thesis will focus on the dilemma of binocular stereo vision camera calibration technology,image preprocessing technology,stereo matching and 3D reconstruction technology and robot path planning technology.The main work of this thesis includes the following aspects.The imaging model of the camera is deeply studied,and the four coordinate systems involved in the camera imaging process are described in detail,and the conversion relationship between them is explained.The calibration method of binocular camera is compared and analyzed.Zhang Zhengyou checkerboard calibration method is determined as the calibration method of this paper,and the calibration algorithm is optimized by maximum likelihood estimation.According to the calibration result,the lens distortion is corrected,and the binocular stereo correction is completed.In image preprocessing,in order to eliminate or reduce the influence of image noise and brightness difference between pairs on stereo matching,mean and Gaussian filtering are used to eliminate salt and pepper noise and Gaussian noise respectively,and histogram equalizes the brightness of the balance map.Differences,image sharpening make the edges of the image clear.The stereo matching algorithms commonly used in comparison are analyzed in detail,and the effect of matching window on stereo matching is studied.For the application scenario of robot path planning,an algorithm based on threshold segmentation to remove ground interference and background is proposed.In the aspect of 3D reconstruction,the Delaunay triangulation technique is deeply studied,and the implementation process of the Lawson algorithm is explained.And based on OpenGL,the 3D reconstruction of the scene surface is completed.A binocular vision hardware system was built and the interface was developed.By designing the robot path planning experiment,not only the previous binocular vision system is verified,but also the practical application effect of binocular stereo vision in robot navigation is expounded.