移动机器人检测识别技术的研究  

A Study on the Detection and Identification Technology of Mobile Robots

作　　者：汤辰 万衡[1] 王凯凯[1] 

机构地区：[1]上海应用技术学院电气与电子工程学院,上海201418

出　　处：《电气自动化》2015年第6期99-101,共3页Electrical Automation

摘　　要：在计算机视觉领域中,对于物体检测识别等技术的广泛应用而受到了大量学者的研究与关注。其中SIFT算法作为一项关键技术依然存在配准实时性较低与耗时问题,针对此问题,采用了局部特征配准的方法,特征点检测的区域缩小了,从而减少了两张图像配准的特征点数目,最终提升了SIFT算法的运行效率。并成功地应用于实验室的服务机器人的平台中,准确性和有效性得到了验证。In the field of computer vision,the widely used object detection and identification technology attracts attention and are researched by a large number of scholars. As a key technique,the SIFT algorithm still has such problems as low instant registration and high timeconsuming. To solve these problems,local feature alignment is adopted to narrow the detection area of the feature point,thus reducing the number of feature points of two-image registration and increasing the operational efficiency of the SIFT algorithm. The said method is successfully applied to the platform of the service robot in the laboratory,and its accuracy and effectiveness are verified.

关 键 词：机器视觉 SIFT算法 局部特征配准 服务机器人 级联分类器 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]