Infrared pedestrian detection method using series combination classification

An effective method using series combination classification was proposed to solve the problem of high false alarm ratio in infrared pedestrian detection. After obtaining the region of interest, two kinds of characters were computed, ratio of width to height and histograms of oriented gradients (HOG). The series combination strategy was used to detect the pedestrian. In the experimental results, the detection ratio reached 90.5% and the false alarm ratio reduced by 43.7%. The results show that the proposed method is effective to detect the infrared pedestrian and reduce the false alarm ratio.