利用遥感图像对震损建筑结构变形检测的识别研究
Recognition of Deformation Detection of Earthquake-damaged Building Structures Using Remote Sensing Images

为识别震后建筑变形损坏状况，提出震损建筑结构变形检测的遥感图像识别分析方法。利用无人机采集震灾区域的遥感图像，将建筑结构变形检测问题转变为构件间坐标测量问题，提取所采集遥感图像中样本矢量点，将其划分为不同种类区域，在此基础上对图像进行聚类分割，以获得震后图像的不同类别建筑结构特征，实现识别不同样本矢量点的地震受灾情况。通过实验分析发现，所提出的图像识别分析方法在一定程度上可以识别出损毁建筑物，但仍需要进一步研究，以提高其识别精度。
To identify deformation and damage to buildings after an earthquake, a remote sensing image recognition and analysis method for deformation detection of earthquake-damaged buildings is proposed. Remote sensing images of earthquake disaster areas were collected by UAV, and the problem of building structure deformation detection was transformed into a problem of coordinate measurement between components. Sample vector points in the collected remote sensing images were extracted and divided into different kinds of regions. On this basis, the images were clustered and segmented to obtain the characteristics of different types of building structures in post-earthquake images, thus identifying the earthquake disaster situation of different sample vector points. Through the experimental analysis, it was found that the proposed image recognition and analysis method can identify the damaged buildings to a certain extent, although further research is needed to improve recognition accuracy