模板自适应支持下的行人检测算法研究  

Research on Pedestrian Detection Algorithm Based on Template Adaptive Support

作　　者：安婧[1] 刘晓娜[1] 赵春艳[1] 

机构地区：[1]兰州文理学院,甘肃兰州730000

出　　处：《兰州文理学院学报（自然科学版）》2017年第3期67-71,共5页Journal of Lanzhou University of Arts and Science(Natural Sciences)

摘　　要：在基于图像的行人检测领域,传统的积分通道特征算法在实际应用中容易出现背景目标的误测与目标丢失等问题.本文针对这些问题,对原有检测算法进行优化改进,并采用基于模板的自适应积分通道特征算法来实现行人检测操作.实验结果表明,通过采用经过优化的预处理操作以及自适应模板,可以实现对丢失目标的有效检测,并对背景干扰进行抑制,从而大幅提高算法的检测处理准确性.In the field of pedestrian detection based on image, the traditional integral channel feature algorithm tends to arouse problems, such as the misdetection of background target and the loss of target. In order to solve these problems, the paper optimizes and improves the original detection algorithm, and adopt the adaptive integral channel feature algorithm based on template to realize the pedestrian detection. The experimental results show that the adoption of the optimized preprocessed operation and adaptive template can effectively detect the loss target, and restrain the background interference, which greatly improves the accuracy of the detection processing of the algorithm.

关 键 词：行人检测 计算机视觉 检测优化 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]