Lane Detection in Critical Shadow Conditions Based on Double A/D Convertors Camera

Lane detection in unstructured environments is the basis for navigation of mobile robots. A method for detecting lane in critical shadow conditions is proposed. Based on the color information of the unstructured lane, an improved region-growing algorithm is employed to segment the image. To enhance the image quality and the accuracy of the algorithm, a double A/D convertors camera is used to recover the color space information of the environments in critical shadow conditions. The results demonstrate that proposed method segments the lane effectively, and is robust against shadows, noises and varied illuminations.