空地一体遥感震灾建筑物多级损毁自动判定研究

After an earthquake occurs, damaged information extraction and damaged grade classification of building provides important decision makers for emergency response actions and later reconstruction. Building damaged detection need all-round and three-dimensional monitoring tools and methods due to the complex structure of building itself and damage information. There are some main problems of existing building damage detection methods, such as lack of the use of three-dimensional features of the building, and the consideration of the damaged grade differentiation, which can influence on the accuracy and reliability of damaged detection. In order to solve these problems, we will study building damaged grade detection using multi-view aerial images(“aerial”) and public sources ground images(“ground”). The main contents of our study include: the description and representation of building damaged factor, automatic register between aerial and ground images, multi-dimension features fusion, and intelligence decision model. The key of our study is the processing integration of remote sensing images, the damaged detection integration, and intelligentialize of damaged grade decision, which can improve the accuracy and automation of building damaged detection, and solve the damage grades decision problem.

震灾建筑物损毁信息提取和等级判定是灾害应急响应和灾后恢复重建的重要决策依据。建筑物自身结构与损毁的复杂性需要全方位、立体的监测手段和方法。针对传统遥感影像建筑物损毁检测无法顾及建筑物结构的立体性和损毁等级的差异性、难以保证检测的精度和可靠性等问题，本课题将基于航空手段获取的多视航空遥感影像（“空”）与地面包括车载/人获取的多源影像（“地”）作为数据源，研究融合空地遥感影像进行震后建筑物多级损毁检测与等级自动判定方法，主要研究内容包括建筑物损毁因子的表征、空地异源影像的自动配准、多维损毁特征的提取以及智能化判定模型。其核心是空地遥感影像处理的一体化、检测处理环节的一体化和损毁等级判定的智能化，从而能有效解决震灾发生后建筑物损毁检测面临的全方位立体空间快速精细检测和多级损毁判定难题。

震灾建筑物损毁信息提取和等级判定是灾害应急响应和灾后恢复重建的重要决策依据。建筑物自身结构与损毁的复杂性需要全方位、立体的监测手段和方法。针对传统方法难以对建筑物进行全方位观测、难以顾及建筑物结构的立体性和损毁等级的差异性等问题，集成利用航空手段（“空”）获取的多视遥感影像和手机/众源获得的地面遥感影像数据（“地”），通过对多级损毁特征描述、空地影像融合、高精度损毁特征提取以及智能化损毁应用等一系列关键问题的研究，实现了自动、精准、高效的建筑物损毁检测。围绕相关模型和方法，开发了典型灾害损毁监测与评估软件、机载在线处理原理样机。相关成果突破了目前建筑物损毁检测方面多角度、全维度的损毁检测瓶颈，克服了理论方法落地应用中实用性、时效性的不足，显著提升了灾损应急服务的智能化与实用化水平。此外，相关成果在近年来灾害损毁评估中发挥了重要作用，并有效支撑了目标损毁评估系统的定型及部署应用。