摘要
[目的]城镇及村落等居民地是人类活动的中心区域,也是遥感影像精准解译的热点和难点。为了完整提取丘陵地区居民地,提出一种多光谱和全色数据协同、分特征提取(SF)的方法。[方法]基于GF1影像对陕西省榆林市横山区丘陵地区的居民地进行自动提取,并对提取结果进行综合研究。[结果]SF相对于常用的基于融合数据提取方法(F)提取正确率提高了19.56%,错提像素个数减少了39.34%,漏提像素个数减少了5.34%。[结论]SF方法不仅可以有效提高丘陵地区居民地提取正确率,而且因为错提量的大大减少,可大量节约后处理工作量。同时,该方法对其他地类的提取也具有良好的借鉴意义。
[Objective]Settlement places,including the urban and village areas,are the central areas of human activities.And they are also the hot and difficult spots for the precise interpretation of remote sensing.In order to extract settlement place in hilly area completely,this paper proposed a new method（SF）that can extract target information from the cooperation of multispectral image and panchromatic image.[Methods]The results had been analyzed comprehensively using GF1 image to extract settlement place in Henshan District Yuling City,Shaanxi Province automatically.[Results]The result showed that the extraction accuracy of SF was improved by 19.56%in comparison with the normal method that extracts target information from fusion data only（F）;the number of incorrect pixels was decreased by 39.34%,and the missed pixel＇s number was reduced by 5.34%.[Conclusion]It can be seen that,SF method not only can effectively improve the extraction accuracy of settlement in hilly region,but also can save post-processing work for the reason that the wrong extraction has been reduced significantly.At the same time,it is also a good reference for the extraction of other land-use type.
出处
《水土保持通报》
CSCD
2017年第6期124-128,共5页
Bulletin of Soil and Water Conservation
基金
国家重大科技专项"高分水利遥感应用示范系统(1期)"(08-Y30B07-9001-13/15)
水利部综合事业局拔尖人才培养专项(2015-132-1)
关键词
居民地
分特征提取
丘陵地区
多阈值分割
settlement place
extraction via different feature
hilly region
multithreshold segmentation
