一种基于结构光双目视觉的特征匹配算法研究  

Research on matching algorithm based on structured light binocular vision feature

作　　者：刘瑜[1] 刘缠牢[1] 苏海[1] 

机构地区：[1]西安工业大学光电工程学院,陕西西安710032

出　　处：《光学仪器》2014年第2期161-166,共6页Optical Instruments

摘　　要：特征点匹配在图像检索、三维测量、模式识别等技术中起着重要的作用。使用MATLAB软件剪切图像并细化线结构光光线条纹。经理论分析SURF算法优缺点,提出了一种基于SURF算法特征点提取的改进算法。用C语言编写改进后的特征提取算法,通过MATLAB软件实验对比两种算法的特征点提取结果并且编写程序实现后期的特征匹配。实验表明:该算法基本满足双目视觉立体匹配的要求,对于线结构光三维测量技术具有重要的理论意义和实用价值。Matching feature points play an important role in image retrieval, dimensional measurement and pattern recognition technology. By using the MATLAB software shear thinning line structured light image and light stripes, and through theoretical analysis of the advantages and disadvantages of SURF algorithm, this article proposes the SURF algorithm which based on feature points extraction algorithm. Through C language improved feature extraction algorithm, MATLAB software experimental comparison of the two algorithms feature point extraction results and preparation program features the late match. Experiments show that the algorithm basically meets the requirements of binocular vision stereo matching, and linear structured light for three-dimensional measurement technique has important theoretical significance and practical value.

关 键 词：结构光 双目视觉 立体匹配算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]