Pedestrian Detection under Low Illumination Environment

Abstract

The moving target detecting is playing a very important role in machine vision, image understanding and other fields. In this paper, the pedestrian detection under the low illumination environment is studied. In order to make the fuzzy pedestrian in the image clearer, the method of combining Gaussian filtering and histogram equalization is used to process the image, which is the base of detecting and extracted Human characteristics. The pedestrian is detected by the method of HOG feature extraction and SVM. To examine the proposed method of the performance detection, experiments have been conducted on the pedestrian image acquired under the low illumination. The results show that the method has clear effect on target image and has achieved the better pedestrian detection under the low illumination environment.