ROAD DIRECTION DETECTION BASED ON GABOR FILTERS AND NEURAL NETWORKS

摘要

This paper proposes a road direction detection system for outdoor autonomous vehicles that must navigate on roads according to visual information. The system consists of three major stages. At the first stage a self-similar family of Gabor kernels are used to extract low-level image features from a road image. Next, unsupervised learning principles with Self-Organizing Maps are used to cluster the Gabor coefficients to higher level image features suitable for road direction detection. At the last stage a multilayer perceptron network trained with a backpropagation algorithm carries out the road direction estimation according to the Gabor coefficients clusters. Several road images taken by a camera mounted on a passenger car are used to test the performance of the system.