Real-time vehicle detection based on Haar features and Pairwise Geometrical Histograms

The Pairwise Geometrical Histograms (PGH) is a generalization of Chain Code Histogram(CCH). It is a powerful shape descriptor that is applied to contours matching which is not affected by rotation. As we know, this method is seldom used for vehicle detection. Feature extraction is a very common and useful method of pattern recognition. In recent years, the Viola and Jones rapid object detection approach became very popular. In this paper, we combine the Haar features and PGH together for vehicle detection. We discuss three methods: the vehicle detection using only Haar features, Haar features combined with hu moments and Haar features combined with PGH. The experimental results show that the last one has the best performance.