Abstract: According to the application of pedestrian detection technology in the intelligent transportation system,in order to improve the efficiency,real-time and accuracy of pedestrian detection method,in this paper,the sparse representation was applied to the feature compression of the image,and a new method of pedestrian detection based on HOG and LTP feature training SVM classifier was proposed.Training SVM classifier for pedestrian detection based on the cha-racteristics of HOG and LTP effectively combines the image gradient feature and texture features and takes advantage of the sparse expression on data compression which can effectively speed up the algorithm.Experimental results show that the proposed algorithm has the advantages of high precision and speed.

Key words: Sparse representation,Pedestrian detection,LTP,HOG,SVM ,Image processing