基于视频监控的震害快速评估与地震应急技术可行性研究

发布时间：2018-02-21 06:56

本文关键词： 视频监控震害快速评估地震应急 GIS 图像处理通信保障　出处：《中国地震局兰州地震研究所》2017年硕士论文　论文类型：学位论文

【摘要】：我国城市地区的人口和经济相对集中,因而也是地震灾害潜在危险性最高的地区。地震防灾、救灾、减灾的主体都在城市,所以提高城市现代化的防震减灾措施,尤其是加强防震减灾信息化管理系统的研制,对于城市的长期稳定发展是必须的。震害评估是防震减灾的必要措施,它是指调查和评价因地震而造成的各类工程结构物的破坏、人员伤亡和经济损失,且多指直接经济损失。本论文试图利用广泛布设于城市道路、公共场所、构(建)筑物内部的视频监控设备所捕捉到的影像资料,通过一些鲁棒性的图像处理方法,结合计算机视觉技术、测震学、地震工程学和通信技术,对震时以构(建)筑物为代表的一系列工程设施进行同震监测,提供实时的数据处理结果和反馈信息,从而在最大程度上避免地震应急指挥的无序性,提高应急工作效率。论文首先介绍了新方法的技术基础,即视频监控影像的计算机图像处理技术。通过人工目视判读和计算机图像分类两种方法结合起来,以此获得监视场景内的目标运动特征及对应参数,从而为地震应急中震害信息的收集工作和后期构(建)筑物的抗震设计提供数据参考。接着集中讨论了震害评估的具体方案设计,针对基于视频监控震害指数和视频监控图像分类两种不同的信息提取方式分别设计震害快速评估方案。最后,文章通过对比几种常见震害评估方法,分析了传统技术在提取震害信息时的不足之处,并适时提出新方法的解决思路。论文还对新技术方案的适用地震强度做了进一步的分析。由于新技术的实现十分依赖稳定的通信保障,因此文章还研究了强震对通信设施的常见破坏情况以及在何等强度下通信设施不被破坏,或者说破坏后依旧能够保障前后方信息交流这一重要内容,并得到了初步的结论——VIII±II度。由此可见,理论上新技术在大震发生后的应用还是非常可靠的。论文最后针对5·12汶川地震主震,进行了应急初期阶段灾区的震害情况快速评估这一正演模型的试算。该模型简化了视频监控影响资料的预处理过程,通过数据插值处理,应用预先设定的算法,验证得出新方法在5·12汶川地震主震这一实例震害快速评估中的应用是可行的这一初步结论。
[Abstract]:The population and economy of urban areas in China are relatively concentrated, so they are also the areas with the highest potential risk of earthquake disasters. The main bodies of earthquake prevention, disaster relief, and disaster reduction are all in cities. In particular, strengthening the development of the information management system for earthquake prevention and disaster reduction is necessary for the long-term and stable development of the city. Earthquake disaster assessment is a necessary measure for earthquake prevention and mitigation, which refers to the investigation and evaluation of the damage to various engineering structures caused by earthquakes. Casualties and economic losses, and more direct economic losses. This paper attempts to make use of the video data captured by the video surveillance equipment which is widely located in urban roads, public places and structures. Through some robust image processing methods, combined with computer vision technology, seismometry, seismic engineering and communication technology, a series of engineering facilities represented by structures (structures) during earthquakes are monitored. Provide real-time data processing results and feedback information, so as to avoid the disorder of earthquake emergency command to the greatest extent, improve the efficiency of emergency response. Firstly, the paper introduces the technical basis of the new method. That is, the computer image processing technology of video surveillance image. By combining two methods, artificial visual interpretation and computer image classification, the target motion characteristics and corresponding parameters in the surveillance scene can be obtained. This provides a data reference for the collection of earthquake disaster information and the seismic design of the post-construction (construction) building in the earthquake emergency response. Then, the concrete scheme design for the earthquake disaster assessment is discussed. Aiming at two different information extraction methods based on video surveillance damage index and video surveillance image classification, the paper designs a fast evaluation scheme of earthquake damage. Finally, this paper compares several common methods of earthquake damage assessment. The shortcomings of traditional technology in extracting earthquake disaster information are analyzed. The paper also makes a further analysis of the suitable seismic intensity of the new technical scheme. Because the realization of the new technology depends on the stable communication guarantee, Therefore, the paper also studies the common damage to communication facilities caused by strong earthquakes and the intensity of which communication facilities are not destroyed, or that the communication facilities can still guarantee the information exchange between the front and rear areas after the destruction. A preliminary conclusion, VIII 卤II degree, has been obtained. It can be seen that the application of the new technique in theory after the occurrence of a large earthquake is very reliable. Finally, the paper aims at the main shock of the 5.12 Wenchuan earthquake. The forward model of rapid assessment of earthquake damage in the early stage of emergency response is calculated. The model simplifies the preprocessing process of the influence data of video surveillance, and applies the pre-set algorithm through data interpolation processing. It is proved that the application of the new method to the rapid assessment of the earthquake damage of the May 12 Wenchuan earthquake is feasible.
【学位授予单位】：中国地震局兰州地震研究所
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P315.9

【相似文献】