Study on Stereo Vision-based Cross-country Obstacle Detection Technology for Intelligent Vehicle

Cross-country intelligent vehicles always work in complicated environments with varying illuminations. The paper presents a new cross-country obstacle detection technology based on stereo vision system. The original images are preprocessed by Gaussian filter and contrast-limited adaptive histogram equalization (CLAHE) method to weaken the effect of noise, light and contrast. Harris corners are located with sub-pixel accurate. To guarantee the overall system real-time performance, feature-based matching techniques are studied and fundamental matrix is calculated based on random sample consensus (RANSAC). Also restrains are studied to eliminate pseudo matching pairs. Then data interpolation is introduced to build elevation maps. Edge extraction and morphological processing are concerned to accomplish obstacle detection. Experiment results for different conditions are presented in support of the obstacle detection technology.