Based on Google Earth free remote sensing image data,this article proposes a highly automated building height extraction method,aiming to solve the problem that the existing methods for obtaining the height of build-ings are high cost,time-consuming and labor-intensive,and have poor timeliness.First,adaptive gamma cor-rection is performed on the original remote sensing image to enhance the image contrast.Then the K-Means unsu-pervised classification method is used to extract the shadows of buildings,and the triangle greenness index is used to identify and eliminate plant pixels that are mistakenly extracted as buildings.Finally,perform post-processing operations to obtain the building shadow extraction results.Based on the geometric relationship between the build-ing height and its shadow,the relevant parameters are deduced using typical known building height data,and the building height information is obtained based on the automatically extracted building shadow length.Experimental results show that the root mean square error of the building height extracted by this method is 0.94m,and the av-erage accuracy is 97.55％.This method can completely extract the shadow of the building,accurately estimate the height of the building,and partly improve automation level.
关键词
建筑物高度/阴影提取/无监督分类/谷歌地球
Key words
building height/shadow extraction/unsupervised classification/Google Earth
维普
万方数据