Multi-target classification method based on the fusion of HOG and SURF features

In order to overcome the problem that HOG+SIFT can not extract the stable features of targets under large illumination changes and the target recognition rate is low, this paper proposes a multi-target classification method based on the fusion of HOG and SURF features. Firstly, the image features are extracted by using the directional gradient histogram (HOG) and the accelerated robust feature (SURF) respectively. Secondly, the two features are fused by the clustering method (K-means), Finally, the support vector machine (SVM) is used to classify the fused features, and the effectiveness of the algorithm is tested on KTH data sets with illumination changes. Experimental results show that the classification method based on HOG+SIFT feature fusion has better robustness and real-time performance. At the same time, aiming at the two main factors (noise and blur) of image degradation in practical applications, the proposed algorithm is tested and analyzed, and the limitations of the algorithm sensitive to noise are obtained.