一种高效且鲁棒的网格化特征提取和匹配方法  

High⁃quality Grid⁃Based Feature Extraction and Matching

作　　者：叶培楚 杨伟钧[2] 曾宪贤 YE Pei-chu;YANG Wei-jun;ZENG Xian-xian(Guangzhou XAG Co.,Ltd,Guangzhou 510663,China;School of Mechanical and Electrical Engineering,Guangzhou City Polytechnic,Guangzhou 510405,China;Guangdong Provincial Key Laboratory of Big Data Computing,The Chinese University of Hong Kong,Shenzhen 518000,China;School of Computer Science,Guangdong Polytechnic Normal University,Guangzhou 510665,China)

机构地区：[1]广州极飞科技股份有限公司,广东广州510663 [2]广州城市职业学院机电工程学院,广东广州510405 [3]广东技术师范大学计算机学院,广东广州510665 [4]香港中文大学(深圳)广东省大数据计算基础理论与方法重点实验室,广东深圳518000

出　　处：《广州城市职业学院学报》2022年第4期61-65,共5页Journal Of Guangzhou City Polytechnic

基　　金：广东省普通高校特色创新类项目“SCR系统便携式检测仪的研究与应用”(编号:2019GKTSCX081);广东省教育科学规划课题“校企深度融合下高职工科双创人才培养模式研究”(编号:2021GXJK673);香港中文大学(深圳)开放课题广东省大数据计算基础理论与方法重点实验室开放课题基金“多源异构融合表征学习理论研究”(编号:B10120210117⁃OF08);广东省普通高校青年创新人才类项目资助“面向多源异构数据精细化融合与协调学习算法研究”(编号:2022KQNCX038)。

摘　　要：在农业三维重建任务中,特征提取和匹配环节是极为重要的一环,对于重建精度和重建效率有较大的影响。文中提出的高质量网格化特征提取和特征匹配方法通过将图像网格化处理,逐块进行特征提取并自适应阈值处理,确保每个图像块中提取到足够多的特征点,并实现特征分布均匀化,提高三维重建算法的精度;通过对最近邻比例法的特征匹配结果进行网格化处理,并基于特征响应值和匹配分数两个维度对匹配对进行筛选,保留所有质量较高且空间分布均匀的匹配对,提高三维重建算法的精度和效率。在现有农业场景中进行实验验证,比较原始特征提取和匹配方法和文中提出的网格化处理的特征提取和匹配方法,证实文中提出的方法可以提取空间分布均匀且质量较高的匹配对,并且可以有效剔除冗余匹配对,提高三维重建匹配效率。In agricultural 3D reconstruction tasks,feature extraction and matching,one of the most important links,exerts a great impact on reconstruction accuracy and reconstruction efficiency.The high⁃quality grid⁃based feature extraction and matching method proposed in this paper ensured that enough feature points were extracted from each image block by gridding the image,extracting features block by block,and adapting threshold processing.In this paper,the feature matching results of the nearest neighbor ratio method were gridded,and the matching pairs were screened based on the two dimensions of feature response value and matching score,and all high⁃quality matching pairs were retained.The matching pairs with uniform spatial distribution improved the accuracy and efficiency of the 3D reconstruction algorithm.Experiments were carried out in the existing agricultural scene to compare the original feature extraction and matching method with that of the grid⁃based feature extraction and matching proposed in this paper.It is confirmed that the method proposed in this paper can propose matching pairs with uniform spatial distribution and high quality,being able to effectively eliminate redundant matching pairs and improve the matching efficiency of 3D reconstruction.

关 键 词：特征提取 特征匹配 网格化 三维重建 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]