【作者】 陈平； 王培珍； 周芳； 王雪峰；

【Author】 CHEN Ping~(1)),WANG Peizhen~(1)),ZHOU Fang~(1)),WANG Xuefeng~(2)) 1)(School of Electrical Engineering & Information,Anhui University of Technology,Ma ’anshan,Anhui,243002) 2)(Research Institute of Forestry Resource Information Technique,Chinese Academy of Forestry,Beijing 100091)

【机构】 安徽工业大学电气信息学院； 中国林业科学研究院资源信息所；

【摘要】 针对苗木图像匹配的特点,采用改进的SIFT算法对自然环境下获取的苗木图像进行了特征量提取;通过添加存入最小优先级队列的限制条件,对现有的BBF匹配算法进行了改进以提高算法的搜索效率;再设置特征点间的几何约束,对由改进后的匹配算法获取的粗匹配结果中的误匹配加以剔除。实验结果表明,提出的改进算法使匹配效率和匹配准确率大为提高。更多还原

【Abstract】 In view of the characteristic of plant image,an improved SIFT(scale invariant feature transform) algorithm is employed to extract features of plantlet images obtained under nature environment.With a constraint of logging minimum priority queue,BBF(Best Bin First) algorithm is modified to improving the search efficiency.Finally,by setting geometrical constraint between feature points,some error matched feature points from modified BBF are eliminated.Experimental results show,with the proposed method,efficiency and accuracy of feature extraction and matching are greatly improved.更多还原