Study, Based On Elliptic Characteristics Of Binocular Visual Space Plane Positioning Methods

As a new multidisciplinary field, the theory research and practical application of the machine vision has made great progress. Especially in recent years, the development of image processing technology and high performance of computer signal processing technology provide favorable conditions for machine vision research and application. Robots based on machine vision have been widely applied. Machine vision has played an important role in location of object. Binocular stereo vision as important branch of machine vision is one of the keys of machine vision research.In this paper, in order to determine the pose of spatial plane based on the binocular stereo vision and ellipse features, the vision model and the ellipse image center extraction algorithm as well as the spatial plane orientation algorithm have been studied and discussed.Firstly, the camera model, the binocular stereo vision measurement principle and its mathematical model are introduced. According to the basic knowledge, the binocular stereo vision model based on laser ellipse features is presented and the relationship between the binocular vision model and the spatial plane is established. It proves that the images of laser ellipse in the left and right imaging plane are elliptical. The equations of ellipses in CCD imaging plane are given by numerical examples.Then, one sub-pixel extraction algorithm about ellipse image center which is called the ellipse fitting method is introduced. The least square ellipse fitting algorithm based on the edge data of laser ellipse is presented. Because the method is an approximate linear fitting process, it has fast calculation speed and high accuracy. The numerical results show that the fitting ellipse equations are the same as the given equations in chapter 2 and the least square ellipse fitting algorithm is effective.Finally, the reasons for the error caused by the use of ellipse center as a match point to acquire depth information are analysed. The search method based on ellipse features and the analytical method based on two ellipse equations are proposed. The numerical results show that the methods are effective and feasible. The search method does not require the matching of feature points and more computation. The calculation error of depth information caused by the center of gravity method is solved. The analytical method also overcomes the defects that need to set feature points in the plane and has less computation than traditional search algorithm.With the development of the research work, the binocular stereo vision system combined with laser token will acquire a satisfied performance. This technology will be widely used in more and more fields.