基于离散分数布朗随机场模型的景象适配性分析方法  

Scene navigability analysis method based on discrete fractional brownian random field model

作　　者：张国忠[1] 王征[2] 蒋秀峰[1] 朱华勇[1] 沈林成[1] 常文森[1] 

机构地区：[1]国防科技大学机电工程与自动化学院自动化研究所,长沙410073 [2]海军装备论证研究中心标准规范研究所,上海200235

出　　处：《宇航学报》2004年第1期19-23,共5页Journal of Astronautics

摘　　要：首先阐述了在自然景象的齐次平稳高斯随机场模型下进行景象适配性分析的不足 ;然后建立了景象的离散分数布朗随机场模型 ,在小波多分辨率分析的基础上提取了分形特征 ;接着推导了景象的分形特征与基准子图的二维相关性之间的关系 ,并提出了基于分形特征的景象适配性分析方法 ;最后通过大量仿真试验验证了分形特征和基准图的正确匹配概率之间的关系 ,对于景象匹配区的选择具有一定的实用价值。The deficiency of scene navigability analysis is examined when scene is modeled by stationary Gauss random field.Then discrete fractional Brownian random field is set up as scene's model.Fractal feature,which differs from the density and edge features,is extracted based on the wavelet representation of the image.The relation between fractal feature and the correlation coefficients of reference sub-map is deduced,and the scene navigability analysis method based on fractal feature is also put forward.Experiments validate the relation between acquisition probability and fractal feature of reference map,which is valuable for the selection of scene matching reference map.

关 键 词：离散分数布朗随机场模型 景象适配性 景象匹配 正确匹配概率 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]