基于GA-BP神经网络的地基干涉雷达监测效能分析  

GB-SAR Monitoring Effectiveness Analysis Using Genetic Algorithm Optimized BP Neural Network

作　　者：杜孙稳[1] 张锦[1] 邓增兵 李静涛 DU Sunwen ZHANG Jin DENG Zengbing LI Jingtao(College of Mining Technology, Taiyuan University of Technology,79 West-Yingze Street, TaJyuan 030024, China Department of Geologic Survey, China Coal Pingshuo Group Co Ltd,Pingshuo Shenghuo District, Shuozhou 036006, China)

机构地区：[1]太原理工大学矿业工程学院,太原市迎泽西大街79号030024 [2]中煤平朔集团有限公司地质测量部,朔州市平朔生活区036006

出　　处：《大地测量与地球动力学》2017年第8期876-880,共5页Journal of Geodesy and Geodynamics

基　　金：国家自然科学基金(41371373);国家863计划(2013AA122301)~~

摘　　要：针对地基干涉雷达露天矿矿区边坡地面灾害监测中影响因素众多且关系复杂的特点,将GA-BP算法应用于GB-SAR形变监测效能分析中。将GB-SAR扫描坡度、扫描坡向以及雷达回波强度3个因子作为神经网络的输入,边坡监测区域实测获取的单位面积上具有形变信息的点个数作为输出,并利用皮尔逊相关系数法分析各影响因素同监测效能的相关性质和相关程度。结果表明,该算法适用于地基干涉雷达的监测效能分析,且具有一定的有效性和优越性。According to the characteristic of the GB-SAR affected by many factors and complex rela- tionship in the open pit mine slope ground disaster monitoring, it used genetic algorithm optimized BP neural network model to analyze the monitoring effectiveness of the GB-SAR. The neural network takes the scan gradient, scan slope direction and the radar echo intensity as the input, takes the ob- tained points number of deformation monitoring as the output, and uses the Pearson correlation coeffi- cient method to analyze the relevant properties and related degree of the influence factors. The results show that the GA-BP algorithm is suitable for monitoring effectiveness analysis of GB-SAR, and it is effective and superior.

关 键 词：监测效能分析 地基干涉雷达 遗传算法 BP神经网络 皮尔逊相关系数法 

分 类 号：P237[天文地球—摄影测量与遥感]