遗传算法和最小二乘匹配相结合的DEM匹配方法  

DEM Matching Method Combining Genetic Algorithm with Least Squares Matching

作　　者：杨容浩[1,2] 岑敏仪[1] 张同刚[1] 杨佳[3] 

机构地区：[1]西南交通大学土木工程学院,四川成都610031 [2]成都理工大学地球科学学院测绘工程系,四川成都610059 [3]成都理工大学核技术与自动化学院,四川成都610059

出　　处：《水土保持通报》2010年第3期128-133,共6页Bulletin of Soil and Water Conservation

基　　金：国家自然科学基金(40271092)

摘　　要：针对传统最小二乘无控制DEM匹配方法拉入范围小的问题,提出了一种遗传算法和最小二乘匹配相结合的无控制DEM匹配方法,为了克服基于传统匹配模型的遗传匹配方法易陷入缩放系数为0的错误全局最优极值处的问题,建立了采用距离等级划分的DEM匹配模型。在此基础上,设计了遗传算法和最小二乘匹配相结合的匹配方法流程。仿真和实际数据实验结果均表明,该方法能够保持最小二乘法匹配精度高和遗传算法拉入范围大的优点,并有较好的稳定性和较高的收敛效率。The traditional methods of DEM matching without control points are based on least squares matching and their pull-in-range are poor.Aimed at this problem,a new method combining genetic algorithm with least squares matching was proposed by this paper.In order to avoid the scaling coefficient converge to the false extreme value 0 easily when using genetic matching method based on traditional matching model,this paper established a new matching model using distance grading.Accordingly,this paper designed the matching procedures of the method combining genetic algorithm with least squares matching.Results from experiments with simulated data and actual data show that the new method proposed by the paper keeps the least squares matching’s merit of higher accuracy and the genetic algorithm’s merit of larger pull-in-range and has better robustness and higher converging efficiency.

关 键 词：无控制DEM匹配 遗传算法 最小二乘匹配 匹配模型 LZD算法 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置] P208[自动化与计算机技术—控制科学与工程]