一种基于特征的自适应地形模型简化方法  

Adaptive terrain model simplification method based on features

作　　者：张慧杰[1,2] 齐峰[3] 王远志[4] 孙红光[1] 欧阳丹彤[2] 

机构地区：[1]东北师范大学计算机学院,长春130117 [2]吉林大学计算机科学与技术学院,长春130012 [3]中国电信股份有限公司吉林分公司网络优化中心,长春130033 [4]安庆师范学院计算机系,安庆246011

出　　处：《仪器仪表学报》2009年第5期1049-1054,共6页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(60773097);国家自然科学青年基金(60603030);东北师范大学自然科学青年基金(20081003);安徽省高校省级自然科学重点项目(KJ2008A18ZC)资助

摘　　要：经研究提出了一种新的根据地形特征自适应确定误差评价函数和表面细分策略的模型简化方法.为了降低特征提取给算法带来的额外开销,建立了细节层次结构,并且提出了与层次结构相结合的特征点选择算法.根据特征点构造了与层次结构中节点相关的特征域,从而限定了特征搜索范围,提高了特征提取的效率.在此基础上建立了特征域的匹配度函数,实现了地势起伏相似区域的合并,最后根据特征域匹配情况自适应选择误差计算方法和阈值.最终达到了提高模型精度和算法效率的双重目的.文中方法在多个基准数据上均进行了实验研究,结果表明与经典层次细分方法相比,算法效率和模型精度均得到显著提高.This paper presents a new terrain model simplification method based on adaptively determining the error function and subdivision strategy for terrain surface, according to the terrain features. In order to decrease the overhead of feature extraction, a hierarchy has been created and a corresponding algorithm of feature point selection has also been put forward. According to these feature points, many feature areas related to the nodes in the hierarchy are produced. Thus, the process of searching terrain features is accelerated greatly. After these adjacent feature areas are merged into one area through their matching degree, an adaptive error evaluation method is applied in the terrain area. As a result, the objective of improving the model precision is achieved. Those algorithms have been applied to many benchmark data, and the results show that our method is more efficient and the result model is of higher precision, compared with the other typical hierarchical subdivision methods.

关 键 词：层次结构 特征点 特征域 区域合并 简化模型 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]