基于高分辨率遥感影像的农村居民点内部用地信息提取研究

近年来,随着农村地区土地利用规划工作的细致深入,对快速获取农村居民点内部用地信息提出了更高的要求。目前,国内外的研究主要集中在对于房屋或道路的提取上,且大多以面向像元的分类方法为主,其结果并不能反映农村居民地内部的整体用地情况且分类精度有待进一步提高。为此,本文提出了一种方法：首先利用大比例尺土地利用现状图提取农村居民点外边界;接着建立起农村居民地遥感分类体系;然后,采用面向对象的分类技术,辅以各种地学知识建立起提取农村居民点内部用地信息的有效模型;最后以武汉市黄陂区的桃园吉地区为研究区域进行了实例研究,验证此方法的有效性。

Extraction of Land-use Information within Rural Residential Area from High-resolution RS Images

Extracting land-use information within rural residential area is one of the major applications in remote sensing today.In this paper,a new method,which is auxiliary land-use knowledge method,is presented for this requirement.With the abundant geographic knowledge in the thematic map,we first propose a simple and effective method to extract rural residential out-border from RS image by overlapping analysis,and take the result as the basic data for further interpretation.Secondly,the object-oriented approach is employed for further classification,whose basic cell isn＇t a single pixel any more,but rather an image object from image segmentation.During the process,land-use knowledge is also taken as auxiliary information to establish class system and class hierarchy,select feature presentation of image objects,and examine classification result.Finally,a high-resolution RS image of Hubei Province is taken as testing data to verify the above method.The experiment results are satisfying：the detailed land-use information is extracted and categories with similar spectrum feature are divided effectively.It is obvious that this method offers a good solution to extract land-use information within rural residential area.