Segmentation of Full Vision Images Based on Colour and Texture Features

Vision guidance is one of the key techniques for autonomous robot navigation, which allows a robot to find a valid path and recognize the environment. This paper analyses and investigates the problems of image-based road detection and understanding. Based on the colour features of the road, an improved region-growing algorithm is used to segment the images. Due to the influence of environment illumination and disturbances, some lane regions may be lost by over segmentation in the image. To improve the accuracy of lane segmentation, wavelet-based texture features and space adjacencies are employed to retrieve the lost lane regions. The experimental results demonstrate that the proposed method can achieve full-image segmentation and is of high precision, robust and reliability for real-time road segmentation and detection.