Robust stereo matching algorithm for advanced driver assistance system
Jamin Koo, Yong-Ho Kim, Sangkeun Lee · 2015
Recently, advanced driver assistance system becomes an important area for automotive industries. Especially, stereo vision-based system has been actively researched. However, in practice, the system has a difficulty in getting correct information because the images from multiple cameras may contain different exposure levels, illumination changes, and lighting geometry differences. To isolate these problems, we present a robust stereo matching algorithm based on color and gradient information for the advanced driver assistance system. Unlike most stereo matching algorithms, the proposed method is robust to the illumination changes and real world's lighting conditions. Experimental result shows that the proposed scheme outperforms other existing methods under the deformed stereo images by various radiometric changes.