Stereo Vision Rectification Based on Epipolar Lines Match and Three Variables Projective Matrix

This paper gives an improved method by increasing precision and decreasing variables for image rectification consistent with Richard I. Hartley's theory and practice of projective rectification (1999), and Loop and Zhang's computing rectifying homographics for stereo vision (1999). This method is based on the camera calibration by Klaus Strobl et al., and DLR CalDc & DLR CalLab, Caltcch University. This paper describes a new discovery of relation between the epipolar lines and projective transformations. Therefore, based on the precise fundamental matrix and match relation of epipolar lines rather than match points, computing the projective transformations matrix is simplified by decreasing the number of unknown variables from the normal eight to three in this paper. The other advantages include the simplicity of the inverse transformation of 2D projective matrix which allows faster resampling in the identification of matched points for the scene reconstruction, and increased rapidity in the process of image rectification.