MBI多特征结合GIS缓冲区的高分影像建筑物变化信息检测

The building change information detection of high resolution image combined with MBI mul-features and GIS buffer

为准确地从高分辨率遥感影像中提取建筑物及其变化信息,基于对建筑物先进行提取再变化检测的思想,在采用改进的分水岭算法进行影像分割基础上,结合形态学建筑物指数(MBI)的多特征支持向量机建筑物提取方法,将GIS缓冲区引入建筑物变化检测.针对配准误差、提取误差和建筑物位移现象引起的误差,构建引入缓冲区的建筑物变化检测流程.以0.2m分辨率航空遥感影像为例,对建筑物变化进行了检测试验,结果表明本文的建筑物变化检测流程能够有效提高建筑物变化检测的准确性.

In order to accurately extract the building and change information from high-resolution remote sensing images, the idea of first extracted building and then detected change was adopted. Based on the image seg mentation with improved watershed algorithm, combinated the SVM method with MBI and multi-feature, the GIS buffer was introduced to change detection. According to the detection error caused by registration error, extraction error and building displacement phenomenon, the building change detection process that introduces GIS buffer was constructed. With 0.2 m resolution aerial remote sensing images case, the building changes detection was tested, the test results showed that building change detection process can effectively improve the accuracy of the building change detection.