A Robust Feature Fusion Method for Camera-Based Highway Guardrail Detection

The guardrail detection system is an important part of environment perception for intelligent vehicles. In this paper, the problem of camera-based highway guardrail detection is considered. A novel feature fusion method, which combines histograms of oriented gradients (HOG) and local binary pattern (LBP), is proposed. Then, principal components analysis (PCA) is utilized to reduce the dimension of proposed feature fusion set. Using the strategy of sliding window, the guardrail is detected by clustering method. Computer simulations show that the proposed method has better performance than other methods.