Development of Automated Damage Detection Technique for Urban Disasters Using High-resolution Satellite Images

利用高分辨率卫星图像开发城市灾害自动损害检测技术

基本信息

批准号：
17310090
负责人：
YAMAZAKI Fumio
金额：
$ 9.92万
依托单位：
Chiba University
依托单位国家：
日本
项目类别：
Grant-in-Aid for Scientific Research (B)
财政年份：
2005
资助国家：
日本
起止时间：
2005 至 2006
项目状态：
已结题

来源：
https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-17310090/
关键词：
High-resolution satellite images Earthquake damage Tsunami The 2004 Indian Ocean tsunami The 2006 Central Java earthquake Optical sensors object-based image segmentation and classification technique Building damage 高解像度衛星地震被害 QuickBir

项目摘要

In this study, automated damage detection technique was developed using high-resolution satellite images. Comparisons between pre- and post-event images were effective to show the affected areas by natural disasters, e.g., earthquakes and tsunami. In addition that, the technique to identify damaged areas only from the post-event images was also considered in this study. The damage detection technique was developed using the high-resolution satellite images captured in the 2004 Indian Ocean earthquake and the 2006 Central Java earthquake, and aerial photographs taken after the 2004 Mid-Niigata earthquake.To detect the affected areas by Indian Ocean tsunami in 2004, the normalized difference vegetation index (NDVI), soil index (NDSI), and water index (NDWI) were compared between the pre- and post-event images. The digital elevation model was also utilized to reveal the damaged areas.As for the 2006 Central Java earthquake, object-based image segmentation and classification technique as well as pixel-based technique have been applied to the satellite images. According to the results, the object-based image segmentation and classification technique was effective to identify the damaged areas especially for high-resolution satellite images.The damaged sections of expressways were detected by image processing using the aerial photographs in the 2004 Mid-Niigata earthquake. The overturning ratio of tombstones in a cemetery was estimated based on an image analysis of aerial photographs.

在这项研究中，利用高分辨率卫星图像开发了自动损伤检测技术。事件前后图像的比较可以有效地显示受自然灾害（例如地震和海啸）影响的区域。此外，本研究还考虑了仅从事件后图像中识别受损区域的技术。损害检测技术是利用2004年印度洋地震和2006年中爪哇地震中捕获的高分辨率卫星图像以及2004年新泻中部地震后拍摄的航空照片开发的。旨在检测2004年印度洋海啸的受影响地区，比较事件前后图像的归一化差异植被指数（NDVI）、土壤指数（NDSI）和水分指数（NDWI）。数字高程模型也被用来揭示受损区域。对于2006年中爪哇地震，基于对象的图像分割和分类技术以及基于像素的技术已应用于卫星图像。根据结果，基于对象的图像分割和分类技术可以有效地识别受损区域，特别是对于高分辨率卫星图像。利用2004年新泻中部地震中的航空照片进行图像处理，检测出高速公路受损路段。根据航空照片的图像分析估算了墓地墓碑的翻转率。