摘要
静止轨道气象卫星影像中,有时诸多区域被多云覆盖,特征点无法识别。为了保证地标匹配精度,需要限制地标选择在无云或少云区域,导致有效地标数量不足、分布不均等问题。提出一种对海陆边界整体轮廓建立多层次格网,通过多层次格网相似度量、逐层逼近实现地标区域影像匹配、有效地标自动选择的方法。首先,介绍以格网表达边界线,使用多层次格网度量影像位置偏差;然后,通过多层次格网自动选择待匹配地标影像,设计并实现基于多层次格网的地标影像匹配算法,自动选择有效地标;最后,分析有效地标区域与格网层次、云量间的关系,以及空间分布特征。实验表明,多层次格网匹配算法能够大大增加全圆盘匹配需要的有效地标,且有效地标在全圆盘中的分布更加均匀。
In geostationary orbit meteorological satellite images,sometimes many areas are covered by cloudy,and the feature points can not be identified.In order to ensure the accuracy of landmarks matching,it is necessary to limit the selection of landmarks in cloudless or less cloudless areas,resulting in insufficient number of effective landmarks and uneven distribution.In this paper,the method of establishing multi-level grids directly for the whole coastline is proposed.The multi-level grid similarity measure and layer-by-layer approximation were used to realize the image matching and automatic selection of effective landmarks.Firstly,the principle of expressing boundary line and measuring image position deviation were introduced based on multi-level grid.Then,the algorithm based on multi-level grid was designed and implemented.Finally,the relationships between the correct matching area and grid level,cloud quantity were analyzed,and spatial distribution was introduced.The experimental results show that the multi-level grid matching algorithm can achieve cloudy image matching,and greatly increase available block images for the full disk matching.The correct matching area distribution is more uniform,which can provide a sufficient number and uniform distribution of block images for geostationary orbit meteorological satellite image matching.
作者
侯绍洋
秦志远
牛磊
张卫国
艾卫涛
高宁
代晗乐
HOU Shaoyang;QIN Zhiyuan;NIU Lei;ZHANG Weiguo;AI Weitao;GAO Ning;DAI Hanyue(School of Surveying and Urban Spatial Information, Henan University of Urban Construction, Pingdingshan 467036, China)
出处
《测绘科学技术学报》
北大核心
2020年第2期160-165,共6页
Journal of Geomatics Science and Technology
基金
国家自然科学基金项目(41701454)
河南城建学院“科研能力提升工程”资助项目(2017YY022)
河南城建学院青年骨干教师资助项目(YCJQNGGJS201806)。
关键词
图像匹配
静止轨道卫星
地标匹配
海陆边界模板
多层次格网
image matching
geostationary orbit satellite
landmark matching
coastline template
multi-level grids
