乡镇级农居建筑物目标判定方法及应用——以辽宁康平张强镇为例  

Objective Determination Method of Rural Residential Buildings at Township Level and its Application——Through Taking Zhangqiang Town,Kangping County,Liaoning Province as an Example

作　　者：李雨佳 张欣然[1] 于浩[1] 田雨佳 王姝婷 Li Yujia;Zhang Xinran;Yu Hao;Tian Yujia;Wang Shuting(Liaoning Earthquake Agency,Shenyang 110006,China)

机构地区：[1]辽宁省地震局,沈阳110006

出　　处：《黑龙江科学》2023年第14期10-13,18,共5页Heilongjiang Science

基　　金：地震应急青年重点任务(CEAEDEM202203)。

摘　　要：农居建筑物是农村地区主要地物类型之一,利用深度学习算法从遥感影像中获取其准确的空间分布信息,能够为地震应急及震害防御工作提供数据支持。以辽宁省沈阳市康平县张强镇为研究区域,利用快鸟卫星获取0.6 m分辨率的张强镇遥感影像,通过PSPNet网络结构及形态学处理完成农居建筑物目标判定实验并应用于建筑物公里网格生成,其建筑物公里网格与实际农居建筑物分布情况一致,能够很好地反映出当地建筑物群体分布特征。As one of the main types of ground objects in rural areas,rural buildings can obtain their accurate spatial distribution information from remote sensing images by using deep learning algorithm,which can provide data support for earthquake emergency and earthquake damage prevention.Through taking Zhangqiang Town,Kangping County,Shenyang City,Liaoning Province as the research area,the remote sensing image of Zhangqiang Town with a resolution of 0.6 m is obtained by using Fast Bird Satellite,and the target judgment experiment of rural buildings is completed through PSPNet network structure and morphological processing,and it is applied to the building kilometer grid generation.The building kilometer grid is consistent with the actual distribution of rural buildings,which can well reflect the distribution characteristics of local building groups.

关 键 词：农居建筑物 PSPNet 形态学处理:建筑物公里网格 

分 类 号：TU198[建筑科学—建筑理论] TP391.41[自动化与计算机技术—计算机应用技术]