最小二乘支持向量机构造海底趋势面研究  

The Study of Constructing Trend Surface by Least Square Support Vector Machine

作　　者：黄贤源[1,2] 翟国君[2] 黄谟涛[2] 隋立芬[1] 

机构地区：[1]解放军信息工程大学测绘学院,河南郑州450052 [2]海军海洋测绘研究所,天津300061

出　　处：《海洋测绘》2010年第3期9-12,共4页Hydrographic Surveying and Charting

基　　金：国家自然科学基金项目(40474007);国家863项目(2007AA12Z326)

摘　　要：为了解决海底趋势面的构造问题,提出了利用最小二乘支持向量机(LS-SVM)重构海底趋势面的方法,并将该方法与趋势面滤波方法进行了分析和比较,同时利用定理证明,趋势面滤波所构造的趋势面只是LS-SVM取特定参数时的解。通过实测算例表明,在构造海底趋势面的过程中,可以通过调整LS-SVM的参数,使其构造的趋势面具有更好的适用性及稳定性,更能反应海底地形的实际情况。In order to solve the problem of trend surface conformation, a new method of constructing trend surface by LS-SVM is presented, then a contrast between this method and the approach of trend surface filter is carried out. The theorem proves that the trend surface filter is the especial result of LS-SVM. The simulated example shows that in the process of constructing trend surface by LS-SVM, the weight parameters could be adjusted, so the trend surface have the property of popular and steady, the results could reflect the sea bed effectively.

关 键 词：最小二乘支持向量机 趋势面滤波 权重系数 

分 类 号：P229.1[天文地球—大地测量学与测量工程]