基于遥感影像的城镇建筑物分类

Classification of Town Buildings Based on Remote Sensing Images

震害损失主要是由建筑物损毁造成的,对城镇建筑物进行有效分类可以做好震害风险防范,通过遥感影像信息提取的方法对建筑物进行分类能提高工作效率。采用多分割图层及多尺度分割技术,利用特征库阈值分类与样本最邻近分类相结合的方法对遥感影像建筑物进行信息提取及分类。分类结果精度评价表明该方法优于利用单一分割图层样本最近邻分类结果,可以用于城镇建筑物分类。根据建筑物分类结果对震害风险进行了划分。

Earthquake damage loss is mainly caused by the damage of buildings,and effective classification of urban buildings can prevent earthquake damage risks.Classification of buildings through remote sensing image information extraction can improve work efficiency.Using multi-segmentation layer and multi-scale segmentation technology,this paper employs the method of feature library threshold classification and sample nearest neighbor classification to extract and classify information from remote sensing image buildings.The accuracy evaluation of classification results shows that the proposed method is superior to the nearest neighbor classification results using a single segmented layer sample,and can be used for urban building classification.According to the results of building classification,the risk of earthquake damage is graded.