本项目针对传统的震损结构快速评估方法中存在的耗时、费力、对专家依赖性强、评估过程存在安全隐患以及资料收集不全的不足，将基于二维图片的三维模型重建技术引入震损结构的快速识别中。针对现有的三维重建算法中存在的问题，重要研究以下内容：二维震损结构图片的疏压缩与选择算法；考虑二维震损结构图片中遮挡效应的特征提取及图像匹配算法；基于B样条曲线的三维震损结构表面重建算法；三维震损结构模型的垂直拼接算法；集成三维图像重构算法并开发手机APP客户端。项目最终目标是为基于三维图像重建技术的震损结构快速评估提供完善的理论、方法和技术。本项目的研究成果将会改变现有的震害调查方式，极大的方便震害调查，节约人力和财力，大大提高震害调查的效率。另外，本课题研究也涉及许多图像处理领域的前沿技术，同时也会促进图像处理相关领域的发展。

This project focuses on the deficiency of traditional earthquake-damage investigation methods of building structures, including time-consuming, exhausting, dependent on experts, potential safety hazard in investigation process, and will introduce 3D model reconstruction based technology into the fast recognition of earthquake damaged structures. The major research contents including: The sparse reconstruction and selection algorithms for 2D seismic damaged pictures; The feature extraction and image matching algorithms considering sheltering; B spline curve based surface reconstruction of 3D model; The vertical merging algorithms for 3D models; The integration of 3D image reconstruction algorithms and development of APP software. The final objection of the project is to provide the sophisticated theory, method and technology for the fast recognition of earthquake damaged structures based on 3D image reconstruction technology. The research results will change the existing way of earthquake damage investigation, provide the great convenience for earthquake damage investigation, save manpower and financial resources, greatly improve the efficiency of the seismic damage investigation. In addition, the research of the project refers to some advanced technologies in image process, so the research of the project can promote the development of image processing related field.