Font Size: a A A

Stereo Matching Algorithm Of Binocular Vision Research

Posted on:2015-07-16Degree:MasterType:Thesis
Country:ChinaCandidate:Q ZhuFull Text:PDF
GTID:2298330452968039Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
Stereo vision technology has been applied to modern people’s normal life, andcontinues to change our standard of living and way of life and quality. Stereomatching technology is an important part of binocular vision, after years of researchand exploration, has obtained the phased research results, but in the featureextraction, stereo matching problem such as the matching strategy still needs to beimproved.This paper deeply studied the basic principle of binocular stereo matchingalgorithm and the method step, based on this knowledge as a concrete realization thegoal of stereo matching feature points, the specific implementation process includesthe following several important steps: camera calibration, feature extraction, stereomatching and depth of information retrieval. In accordance with the above steps toimplement the system in this paper, first of all, according to the principle of thecamera calibration, combining this article selects the calibration method ofcalibration method (method zhang) the camera calibration and obtain the internal andexternal parameters; Then implement and PCA-SIFT algorithm is studied, a usingCanny edge detection algorithm is proposed to replace the original owners ofcurvature to remove edge response of improved PCA-SIFT algorithm; Then thegravitational search algorithm in stereo matching, in under the condition of highreal-time demand, according to the gray level of image information to extract thefeature points of partition, will have the same feature points are divided into thesame area of grayscale, recycling the search strategy for parallel search the way to improve the original gravity gravitational search algorithm; Finally,3d coordinatecalculation principle to build hardware platform, finally restore the depth of theobject information.By the final experiment shows that improved PCA based on Canny operator,SIFT feature extraction algorithm to enhance the stability of the extraction of featurepoints; Will improve the gravitational search algorithm as the search strategy tospeed up the match speed, reduce the matching time. Satisfy the requirement of highreal-time and improves the matching accuracy.
Keywords/Search Tags:Binocular stereo, PCA-SIFT algorithm, Feature extraction, Gravitational search algorithm, Stereo matching
PDF Full Text Request
Related items