基于多源遥感数据的农村居民点用地提取技术

Extraction Technique of Rural Residential Land Based on Multi-source Remote Sensing Data

摘要:

在分析和改进传统的建设用地、不透水面遥感分类方法的基础上，针对农村居民点用地的特点，提出了一种结合面向对象分类技术、多源、多时相遥感数据数据综合分析技术，借助先验知识，通过遥感手段与GIS技术相结合进行农村居民点用地分类提取的新方法。使用北京市遥感数据进行了实验，自动分类结果准确率高。另外，通过对图像分割后得到的图像对象进行多边形编辑，可以进一步提高准确率，从而满足高精度研究和监测的需要。

Abstract:

In this article，based on the analysis and improvement of traditional method of remote sensing classification for construction land and impervious surface, according to the characteristics of rural residential land, combining with comprehensive analysis technique of object-oriented classification，multi-source and multi-temporal remote sensing data，by using prior knowledge，through a combination of remote sensing and GIS technology，a new method has put forward to the extraction of rural residential land classification. An experiment has been carried out by using remote sensing data in Beijing，it is showed that the result has high accuracy on automatic classification. In addition, it can improve accuracy and satisfy the demand of high-precision studying and monitoring by editing polygon obtained after partitioning images.