Improved adaptive stereo matching using double dynamic programming with correlation of row and column

In a variety of stereo matching algorithms using dynamic programming algorithm can effectively improve the speed and accuracy with good real-time performance and easy realization. In this paper, with advantages of dynamic programming, we presents a stereo matching algorithm using double dynamic programming with correlation of row and column, which uses improved adaptive support-weight approach and energy minimization model to solve the optimization problem. In the solving process, different constructed values of data items given based on characters of variation in the column direction using row dynamic programming and the values of smoothness item in the column direction, to reduce apparent defective stripes, are used to get the final result using column dynamic programming. Experimental results used Middlebury database show that the proposed algorithm can reduce the overall matching error rates and apparent defective stripes and get a better stereo matching effect.