Abstract: Autonomous navigation applications demand sensors with a low sample time to be able to increase speed. Stereo vision algorithms that produce a dense disparity map prsent a slow time response (half a minute per frame at high resolution). We have developed a new cost function for dynamic programming stereo algorithms, capable to deliver dense disparity maps for single, high-resolution scanlines at high speed (40 ms/line), even for wide disparity ranges (>100). We have tested the algorithm with both synthetic and real image, and we have compared its practical performance with other dynamic programming algorithms. Our cost function is based on a weighted sum of the squared intensity errors. Weight factors are based on gradient values. The occlusion cost is not constant for the whole image, instead we modify it depending on gradient values of matched points. In this paper we present this cost function, and compare its performance for autonomous navigation applications with other dynamic programming solutions.