Research Of Depth Information Restoration Based On Stereo Vision

As a new field based on various interdisciplinary,the theory research and practicalapplication of the Computer Vision make great progress. Especially in recent years, thedevelopment of image processing technology and high performance computer coming into beingprovide favorable conditions for Computer Vision research and application. So, the furtherresearch to Computer Vision has much theory meaning and practicality value. Stereo Vision is to restore the depth information of an object from some of itstwo-dimensional images. Generally, one would like to restore the depth information with onlytwo different view points. The fundamental principle of Computer Vision with dual-camera has astrong resemblance to the visual perception procedure of mankind. We can get different imageswith two different view points and then calculate the position deviation of images pixels basedon triangular measure. Finally, we can attain the depth information of the object.The integrated binocular stereo vision system generally includes image acquisition, cameracalibration, features extraction, stereo matching, depth restoration and depth interpolation, etc.The image preprocessing, the features extraction and the stereo matching is mainly discussed inthe paper.First, we elaborate the fundamental principle of measure distance in stereo vision based onthe camera model. Second, the paper gives the elaborate analysis on image preprocessing andfeatures extraction methods, and verifies the algorithms by examples. Furthermore, itsummarizes the content of stereo matching and method of stereo matching. We submit thetemplate matching algorithm based on features constraint based on the edge features and matchalgorithm by features of the roads' images, use the template matching algorithm to matches theroad roads and get good results. Finally, depth information is restored with parallax throughstereo matching.