基于SIFT检测与匹配算法的奶牛形体图像技术研究  

STUDY OF THE TECHNOLOGY BASED ON SIFT DETECTION AND MATCHING ALGORITHM IN THE COW 'S CONFORMATION IMAGES

作　　者：王立中[1] 麻硕士[1] 薛河儒[1] 侯振杰[1] 

机构地区：[1]内蒙古农业大学,呼和浩特010018

出　　处：《内蒙古农业大学学报（自然科学版）》2009年第1期238-242,共5页Journal of Inner Mongolia Agricultural University(Natural Science Edition)

基　　金：教育部"春晖计划"项目(Z2007-1-01010)

摘　　要：使用普通数码相机拍摄奶牛形体图像,应用SIFT特征点检测与匹配算法进行特征点检测、匹配、基础矩阵和单应投影矩阵的计算。实验结果表明,该方法对图像采集的要求低,能适应野外作业,检测的特征点数量大,精确度高,匹配结果准确。The Cow＇s Conformation Images is shot by the common digital camera, and SIFT feature detecting and matching algorithm is applied to feature points detecting, matching, and calculating fundamental matrix and homography projection matrix. Experiments demonstrate that this method has low demand to shooting images and can adapt field task, and has a good performance of locating large numbers of feature points accurately, and matching feature points exactly.

关 键 词：奶牛体型 特征点检测 SIFT算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]