摘要
多源高分辨率遥感影像的自动匹配是研究领域的一个技术难题。针对目前尺度不变特征变换(scaleinvariant feature transformation,SIFT)匹配算法多存在计算量大、耗时长等问题,在SIFT匹配算法的基础上提出了一种改进算法。首先,根据遥感影像直方图大致呈瘦钟形和二次多项式拟合瘦钟形较为理想的特性,采用二次多项式系数来表示关键点及周围像素的整体特征;其次,利用关键点初步筛选占比函数和二次多项式系数对待匹配图像中所有关键点进行粗匹配;最后,利用关键点局部特征的128维向量实现快速精匹配。实验分析表明,采用整体和局部特征相结合的关键点快速匹配算法能够在保证匹配精度的前提下,提高算法自动匹配的效率。
The automatic matching of multi-source high-resolution remote sensing images is a technical challenge in the research field.The current scale-invariant feature transformation(SIFT)matching algorithms often have problems such as large amount of calculation and long time-consuming.This paper proposes an improved algorithm based on SIFT matching algorithm.Firstly,it uses the coefficients of quadratic polynomials to represent the overall characteristics of the key points and the surrounding pixels,according that remote sensing image histogram is roughly thin bell-shaped and the quadratic polynomial can be appropriate for fitting the thin bell shape.Secondly,initially filter function and the second-order polynomial coefficient are used to roughly match all the key points in the matching image.Finally,fast precision matching is achieved using 128-dimensional vectors of local features of key points.Experimental analysis shows that the key point fast matching algorithm combining the whole and local features can improve the efficiency of the automatic matching algorithm under the premise of ensuring the matching accuracy.
作者
薛理
杨树文
马吉晶
刘燕
XUE Li;YANG Shuwen;MA Jijing;LIU Yan(Faculty of Geomatics,Lanzhou Jiaotong University,Lanzhou 730070,China;Gansu Provincial Engineering Laboratory for National Geographic State Monitoring,Lanzhou 730070,China)
出处
《遥感信息》
CSCD
北大核心
2019年第4期54-61,共8页
Remote Sensing Information
基金
国家重点研发计划(2017YFB0504201)
国家自然科学基金(41761082)
兰州市人才创新创业项目(2015-RC-28)
兰州交通大学优秀平台支持(201806)
作者简介
薛理(1993—),男,硕士研究生,主要研究方向为遥感数据处理与分析。E-mail:1151617653@qq.com;通信作者:杨树文(1975-),男,教授,主要研究方向为遥感信息识别与提取。E-mail:ysw040966@163.com.
