融合颜色和深度信息的运动目标提取方法  

A Moving Object Detection Method Combining Color and Depth data

作　　者：胡涛[1,2,3,4] 朱欣焰 呙维[1,2] 张发明 HU Tao;ZHU Xinyan;GUO Wei;ZHANG Faming(State Key Laboratory of Information Engineering in Surveying,Mapping and Remote Sensing,Wuhan University,Wuhan 430079,China;Collaborative Innovation Center of Geospatial Technology,Wuhan 430079,China;School of Resource and Environmental Sciences,Wuhan Umversity,Wuhan 430079,China;School of Library and Information Science,Kent State University,Kent 44240,Ohio,USA)

机构地区：[1]武汉大学测绘遥感信息工程国家重点实验室,湖北武汉430079 [2]地球空间信息技术协同创新中心,湖北武汉430079 [3]武汉大学资源与环境科学学院,湖北武汉430079 [4]肯特州立大学信息科学学院,美国俄亥俄州肯特44240

出　　处：《武汉大学学报（信息科学版）》2019年第2期276-282,共7页Geomatics and Information Science of Wuhan University

基　　金：国家自然科学基金(41301517);国家重点研发计划(2016YFB0502204);中央高校基金(413000010);测绘遥感信息工程国家重点实验室开放研究基金((16)重03)~~

摘　　要：行人检测是计算机视觉、智能交通等领域研究的热点与难点,基于深度传感器对室内复杂场景下的行人检测展开研究。目前,基于颜色与深度数据的目标检测方法主要包括基于背景学习的方法和基于特征检测算子的方法,前者依赖于视频序列头几十帧的背景知识,帧的数量决定检测质量;后者存在计算量大的问题,训练样本的不足也会影响行人检测结果。因此,深入分析了复杂场景特征,融合颜色和深度信息,提出了RGBD+ViBe(visual background extractor)背景剔除方法,实现前景运动目标的准确提取。实验结果表明,提出的RGBD+ViBe方法在前景运动目标检测准确率方面要明显高于仅考虑颜色或深度信息方法以及RGBD+MoG(model of Gaussian)方法。Pedestrian detection is a hot topic and difficult problem in areas like computer vision and intelligent traffic. This paper researches on pedestrian detection in complex indoor space using depth sensors. In recent years, target detection methods based on RGB-Depth(RGBD) data mainly include background learning method and feature detecting operator method. However, the former method depends on the background knowledge of first tens of frames, and the number of frames decides the final detection accuracy. The latter takes plenty of time for computing, and being lacks of training samples may influence the detection result. Thus, this paper analyzes the complex scene features and integrates the color and depth information, and proposes a RGBD+ViBe(visual background extractor) background elimination method. The experiment results indicate that the detection accuracy of the proposed method is higher than the methods which only consider color or depth information and the RGBD+MoG method in foreground extraction.

关 键 词：深度传感器 RGBD 运动目标提取 行人检测 ViBe 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] P208[自动化与计算机技术—计算机科学与技术]