Matching Algorithm Based On Energy Minimization Of The Global Three-dimensional Study

Visual stereo matching is one of the fundamental and significant problems in the study of the computer vision and contactless measurements. This technique makes it possible to reproduce a three-dimensional stereo by getting the distance information through pixels. On the hand, it is the most difficult problem to be solved completely. The most important and hard scenario of computation stereo is computing correspondence(i.e. obtaining disparity map). How to obtain disparity map, which meets the requirement of accuracy is focused on by most stereovision researchers throughout the world.Matching is performed directly with the image intensity in the area-based correlation without using the color information, but in this paper, all algorithms use this information. Algorithm based on global energy minimization is one of the best categories of stereo matching algorithm. After reading many papers about stereo correspondence and studying the dynamic programming and belief propagation algorithm, we propose some new ideas about these.First, we develop the dynamic programming algorithms. Because of the streaks in one scanline by mismatch, this paper proposed a new method to how to get the ground control points by using the edges information, and using these points to control the paths to solve this problem.Second, to improve the accuracy of the disparity map by removing the mismatches caused by both occlusions and false targets, we proposed a new stereo algorithm including partial matching of segmented image and particle swarm optimization of the rest . The experiments result shows that the algorithm is efficient.