融合特征点与环状域检测的水下声呐目标匹配  

Underwater sonar target matching between fusion feature points and ring domain detection

作　　者：李鹏[1,2] 马味敏 Li Peng;Ma Weimin(Jiangsu Key Laboratory of Meteorological Observation and Information Procession,Nanjing University of Information Science and Technology,Nanjing 210044,China;Jiangsu Meteorological Sensor Network Technology Engineering Center,Nanjing University of Information Science and Technology,Nanjing 210044,China)

机构地区：[1]南京信息工程大学江苏省气象探测与信息处理重点实验室,南京210044 [2]南京信息工程大学江苏省气象传感网络技术工程中心,南京210044

出　　处：《电子测量与仪器学报》2019年第11期120-127,共8页Journal of Electronic Measurement and Instrumentation

基　　金：江苏省第11批六大高峰人才项目(2014-XXRJ-006);江苏省重点研发计划社会发展项目(BE201569);国家自然科学基金(41075115)资助项目

摘　　要：由于水下环境复杂,声呐成像背景噪声大,水下目标匹配存在一定难度,提出了一种SURF算法结合恒虚警算法来提取目标区域特征点对,其次通过环状域检测来进一步寻找最优匹配点对,最后计算最优特征点对之间的几何关系确定尺度变换,旋转变换的大小,对待匹配图像进行旋转校正后可以更精确匹配地目标区域。实验与分析结果表明,计算模板图像与待匹配的目标区域之间的匹配重叠率,可以通过受试者工作曲线直观的看出该算法相比于BBS算法和改进的BBS算法的鲁棒性和抗噪性能更好,平均匹配准确率分别提高了32.59%和16.57%。Due to the complex underwater environment, the background noise of the sonar imaging is large, and the underwater target matching is difficult. In this paper, a SURF algorithm is combined with a constant false alarm algorithm to extract the feature points of the target region. Secondly, the annular region detection is used to further find the optimal matching point pairs. Finally, the geometric relationship between the optimal feature point pairs is determined. The size of the rotation transform, the rotation correction of the image to be matched can more accurately match the target area. The experimental and analytical results show that the matching overlap ratio between the reference image and the target region can be used to visually see the robustness of the proposed algorithm compared to the BBS algorithm and the improved BBS algorithm. The anti-noise performance is better, and the average matching accuracy increased by 32.59% and 16.57% respectively.

关 键 词：SURF算法 恒虚警算法 最优特征点对 环状域 受试者工作曲线 匹配重叠率 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]