Steering wheel boundary feature extraction based on contour fusion for long-distance bus

Abnormal driving behavior monitoring and precaution is one of the important parts of advanced driver assistance systems (ADAS). This paper developed a boundary feature extraction algorithm based on contour fusion for steering wheel. The image pre-processed method including median filter, histogram equalization and gray level stretching were used. Boundary feature of steering wheel was extracted by using adapted Sobel operator and image was binarized with Otsu algorithm base on subimage. Contour fusion method was utilized to detect whole boundary of steering wheel. Experimental results show that the algorithm has good adaptability to illumination changing, better robustness and less computation, which lay a stable basis on intelligent analysis of abnormal driving behavior for long-distance bus.