>> 自然科学版期刊 >> 2014年04期 >> 正文
基于双目视觉的目标识别与定位
供稿: 高如新;王俊孟 时间: 2018-11-21 次数:

作者:高如新;王俊孟

作者单位:河南理工大学电气工程与自动化学院

摘要:给出一种双目视觉系统结合SURF(Speeded Up Robust Feature)特征的目标识别与定位方法,分析了采用SURF特征实现目标识别的方法,利用双目视觉的原理求取识别出的目标物体的三维坐标信息,从而实现目标物体的识别与定位,与传统的基于SIFT(Scale Invariant Feature Transform)特征或基于颜色形状特征的目标识别与定位方法相比,该方法更具有鲁棒性和实时性.实验结果证明了该方法进行目标识别与定位的可行性,具有一定的应用价值.

基金:河南理工大学博士基金资助项目(B2010-17);

关键词:SURF特征匹配;双目视觉;目标识别与定位;

DOI:10.16186/j.cnki.1673-9787.2014.04.023

分类号:TP391.41

Abstract:A method for target identification and orientation based on the combination of SURF feature and binocular vision is presented. By means of the principle of binocular vision, the three-dimensional coordinate information of an object is obtained, which had been identified by use of SURF feature. Thusly, the target recognition and location are achieved. Compared with the traditional method, such as ones using SIFT feature, color or appearance feature, the method proposed is more robust and has a high-performance. The feasibility and application value of this method used for target recognition and positioning has been proved according to experimental results.

最近更新