文献类型：期刊文献

英文题名：BD_CNN based pedestrians and vehicle recognition in video surveillance

作者：Zhang, Xueqin[1]; Fang, Ting[1]; Gu, Qiuchen[2]

机构：[1] Institute of Information Science and Technology, East China University of Science and Technology, Shanghai, China; [2] Institute of Business, East China University of Science and Technology, Shanghai, China

年份：2017

卷号：685

起止页码：185

外文期刊名：Communications in Computer and Information Science

收录：EI(收录号：20171203487850)

语种：英文

外文关键词：Vehicles - Pedestrian safety - Object detection - Convolution - Security systems

摘要：Intelligent video analysis technology has been widely used to detect moving vehicles and pedestrians in huge volume video record. In order to find a unified solution to moving vehicles and pedestrians detection in videos, BD_CNN (Background Difference_Convolutional Neural Network) algorithm is proposed in this paper, which applies background difference technology to recognize moving objects, adopts convolutional neural network algorithm to extract features of vehicle and pedestrian automatically, and builds unified vehicle and pedestrian classifier. The multi-scale detection method and reasonable decision mechanics are designed to improve the detection accuracy. Experimental results validate the algorithm’s availability and efficiency in the moving objects detection. ? Springer Nature Singapore Pte Ltd. 2017.