Traffic congestion judgment based on linear spatial pyramid matching using sparse coding

Traffic congestion judgement is a frequently addressed problem in intelligent transportation system. In this paper, a judgement algorithm for identifying the occurring traffic congestion of vehicles is experimentally designed. This algorithm extracts the SIFT features from an image containing vehicles using the linear spatial pyramid matching using sparse coding (ScSPM), then judges weather the congestion is occurring or not. A number of experiments are conducted and compared in this paper to evaluate this algorithm. However, in order to compare the performance of the proposed ScSPM operator with some others, two classic classification algorithms SVM and SPM are used as the references. Through these comparisons show that the judgement algorithm based on ScSPM is efficient and performs better than the other two.