The Design And Implementation Of Pedestrian Detector Based On Multi-feature Fusion

With the development of human society and the rapid increase of population density, some problems such as how to control traffic accurately and effectively, how to make the public monitoring more simpler and effective, how to find the people who is interested fastly and accurately become more significance. Pedestrian detection is pedestrian detection in image or video monitoring. It is the most basic and most important step in some applications such as intelligent transportation, intelligent video surveillance, pedestrian flow diction in various scenarios. Due to the appearance of the pedestrians are susceptible to scale, clothes, movement, light and shade, the pedestrian detection is quite difficult. Therefore, the pedestrian detection become the hot and difficult research point in the field of computer vision. And pedestrian detection is facing two major problems which are the accuracy and real-time performance.This paper proposes a pedestrian detection algorithm which mainly using HOG feature build a detector of whole human body and using Edgelet feature build some detector of human parts. And then this paper use joint probability to combine all the detector to detect pedestrian. This paper also uses variance filter and integral figure to improve the speed of operation.After experimental verification, the pedestrian detector this paper designed can response to some shade of pedestrian and has good real-time performance...