基于0.2m分辨率遥感影像的震害信息提取仿真

Simulation of Earthquake Damage Information Extraction Based on 0.2m Resolution Remote Sensing Image

现有的震害信息提取方法忽略了对遥感图像的几何校正,且遥感图像分辨率的合理性较差,导致提取的影像信息精度不高。于是提出基于0.2m分辨率遥感影像的震害信息提取方法。结合地物光谱的成像与解译过程采集遥感图像。利用共线方程对图像完成几何校正。采用数字元纠正算法获取地面与像点坐标,对所有像素正射校正,再通过图像配准,增强图像质量。设置分割参数,计算空间地物的异质度与相关权重,以像素图斑为中心做图像分割,划分不同特征目标。分析各类震害在图像中表现出的特征,采用卷积神经网络算法,提取出最终震害信息。仿真结果表明,0.2分辨率采集的遥感图像较为清晰,能够获得更加精准的震害信息,为震后救援提供决策依据。

In the existing methods, the geometric correction for remote sensing images is always ignored. The rationality of image resolution is also not ideal. Therefore, a method to extract seismic disaster information based on 0.2 m resolution remote sensing images was presented. Firstly, remote sensing images were collected in combination with the imaging and interpretation process of ground-object spectra. Secondly, collinearity equations were used to complete the geometric correction for images. Then the digital element correction algorithm was adopted to obtain the coordinates of ground and image points. After all pixels were corrected by ortho-rectification, the image registration was used to enhance the image quality. Thirdly, the segmentation parameters were set to calculate the heterogeneity and relevant weights of the spatial objects. After that, the pixel spot was used as the center for image segmentation to divide different feature targets. Furthermore, the characteristics of all seismic disasters in the image were analyzed. Finally, the convolutional neural network algorithm was used to extract the seismic disaster information. Simulation results show that the remote sensing image with a resolution of 0.2 m is clearer, and the seismic disaster information is more accurate. This method can provide a decision-making basis for post-earthquake rescue.