双目立体视觉中的一种运动目标检测算法  

A Moving Object Detection Algorithm in Binocular Stereo Vision

作　　者：杨丽娟[1] 李伟红[1] 赵丽艳[1] 

机构地区：[1]北华航天工业学院计算机科学与工程系,河北廊坊065000

出　　处：《北华航天工业学院学报》2008年第4期4-7,共4页Journal of North China Institute of Aerospace Engineering

摘　　要：运动目标的检测是计算机视觉研究领域中的一个重要组成部分,传统的基于单目视觉的目标检测算法要么受环境因素的限制较大,要么收敛速度较慢,还难以解决多个运动目标部分遮挡的问题。针对以上存在的问题,本文在目标检测中采用了立体视觉方法。用双目摄像机对图像进行拍摄获取立体图像序列对,然后对立体图像对运用立体视觉方法进行分析。实验表明该方法能正确检测到运动目标,不受光线变化和阴影的干扰,并且在运动目标发生部分遮挡时仍能正确区分各目标。Moving object detection is an important part of computer vision research area. The traditional detection algorithms based on monocular computer vision are restricted by environment and convergent speed and unable to solve the shelter problem between multiobjects. To solve these problems a stereo vision algorithm is proposed. In this algorithm, a series of stereo images using a binocular camera are obtained firstly, and then these obtained images are analyzed using a stereo vision method. The experiment shows that this algorithm can detect the moving object correctly without interference with light shade and the shelter problem

关 键 词：运动目标检测 双目立体视觉 计算机视觉 立体图像 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]