Error Reduction by Artificial Intelligence Based Search and Bidirectional Scan in Dynamic Programming in Stereo Vision

The determination of three dimensional information about the scene from two or more two dimensional views of the same scene has been a problem widely discussed in study of computer vision. Stereovision uses two cameras which capture a pair of image locations and then a single three dimensional physical location is derived by exploiting a number of constraints. For the purpose of exploiting these constraints local and global methods are used that attempts to match pixels in one image with their corresponding pixels in the other image. In our paper we use the global method:Dynamic Programming to find the correct match between the two images and also to improve the efficiency we use the bidirectional matching in which dynamic programming is applied from left to right and then from right to left to avoid occlusion and A* search for increasing the quality of the results is to use an optimal searching methodology for the minimal cost path.