基于高分辨率遥感影像的建筑物提取方法研究——基于ENVI平台的高分影像建筑物提取技术实证

Research on Building Extraction Method Based on High-Resolution Remote Sensing Imagery—Empirical Study on Building Extraction from High-Resolution Imagery Based on the ENVI Platform

摘要: 在城市发展中，建筑物具有关键作用，在高分辨率遥感影像上也占据显著地位。随着电子地图与GIS技术的进步，建筑物精确提取的需求日益增强。基于高分影像提取的建筑物数据，可用于面积估算、城市布局分析，并为城市规划、灾害监测和生态评估等提供支持。本文基于高分辨率遥感影像，开展建筑物提取研究。通过梳理研究背景与近期文献，采用监督分类和面向对象分类等方法进行提取，并开展分类后处理及精度验证。结果表明，在本研究区域中，基于规则的面向对象分类方法效果最优，总体分类精度达99.12%，Kappa系数为0.9826，错分误差为0.54%，漏分误差为0.47%，制图精度为99.53%，用户精度为99.46%。研究也发现，建筑物的波谱、形状和大小，以及影像波段与太阳直射角引起的色差与阴影，均会对提取精度产生影响。

Abstract: In urban development, buildings play a key role and occupy a significant position in high-resolution remote sensing images. With advances in electronic maps and GIS technology, there is an increasing demand for accurate building extraction. Building data extracted from high-resolution images can be used for area estimation, urban layout analysis, and provide support for urban planning, disaster monitoring, and ecological assessment. This paper conducts research on building extraction based on high-resolution remote sensing images. By reviewing the research background and recent literature, extraction is performed using methods such as supervised classification and object-oriented classification, followed by post-classification processing and accuracy verification. The results show that in the study area, the rule-based object-oriented classification method achieves the best performance, with an overall classification accuracy of 99.12%, a Kappa coefficient of 0.9826, commission error of 0.54%, omission error of 0.47%, map accuracy of 99.53%, and user accuracy of 99.46%. The study also finds that building spectra, shape, and size, as well as color differences and shadows caused by image bands and solar incidence angles, all affect extraction accuracy.