Research on AD Stereo Matching Algorithm Based on Improved Census Transform
In stereo matching algorithm,aiming at the problems of Census algorithm for image depth discontinuity,weak texture area matching accuracy is low,easy to be disturbed by noise,an improved Census transform combined with adaptive window method is proposed,and gradient information is used to realize adaptive window selection.Moreover,the Absolute Difference(AD)calculation and the improved Census transform were integrated to improve the accuracy and anti-noise ability of the algorithm.In the cost aggregation stage,different adaptive thresholds were provided through different adaptive Windows.Finally,the winner-take-all algorithm(WTA)was used for parallax optimization.And then we get our final parallax map.Experiments on standard images provided by Middlebury Platform show that the improved algorithm has lower mismatching rate and better anti-noise ability.
NETL
NSTL
万方数据
维普
