A video-based indirect sensing procedure for dynamic identification purposes is presented. To overcome major limitations of video-based methods in real on-site measurements, a novel three step pre-modification, magnification, post-modification process is developed. This process includes revision of the initial input video record in order to delete disturbing objects, utilizing a magnification method to filter the frequency content of the monitored motion and using a revision step for elimination of noises generated during magnification process. Finally, a set of digital signal and image processing analyses are performed on the modified video using virtual visual sensor technology. Based on the results of this research, motion signals of the monitored object are detected. The proposed approach has been used for identification of dynamic characteristics of two historic masonry minarets in Istanbul. Results shows that the proposed procedure is able to assess the dynamic characteristics of the monitored structure with a high-level of accuracy.

提出了一种用于动态识别目的的基于视频的间接传感程序。为了克服基于视频的方法在实际现场测量中的主要局限性，开发了一种新颖的三步预修改、放大、修改后过程。该过程包括修改初始输入视频记录以删除干扰对象，利用放大方法过滤被监控运动的频率内容，并使用修改步骤消除放大过程中产生的噪声。最后，使用虚拟视觉传感器技术对修改后的视频进行一组数字信号和图像处理分析。根据这项研究的结果，检测被监控对象的运动信号。所提出的方法已用于识别伊斯坦布尔两座历史悠久的砖石尖塔的动态特征。结果表明，所提出的程序能够以高精度评估被监测结构的动态特性。