论文部分内容阅读
当前,经典的2维遥感变化检测方法不再适用于地质灾害引起的地表3维变化。此外,一般植被提取依赖近红外波段,而近红外影像未必能获取。因此,针对2010年和2011年汶川映秀地区的可见光航空立体像对,提出了一种检测植被和3维地形变化的方法。首先,生成研究区两个不同时期的DEM和DOM并完成配准,再利用CIE Lab色彩空间和Otsu分割算法对可见光DOM进行植被变化检测。然后,基于概率统计理论对新旧时期差分DEM提出了3维变化检测的自适应阈值确定方法,在高概率置信域条件下提取出地质灾害的高危区域。最后,对高危区域运用离散化积分方法估算出土方量的3维变化量。实验结果表明了所提出方法的可行性,有效性和实用性。本文不仅将常规遥感的2维变化检测升华至3维空间,而且对变化的3维地形进行了定量估计,可应用于地质灾害的遥感动态监测和评估。
At present, the classic 2-D remote sensing change detection method is no longer suitable for 3-D surface changes caused by geological disasters. In addition, the general vegetation extraction depends on the near-infrared band, and near infrared images may not be able to obtain. Therefore, aiming at the visible light aerial stereo pair in Yingchuan area of Wenchuan in 2010 and 2011, a method to detect vegetation and 3-D topographic change is proposed. First, DEM and DOM were generated and registered in two different periods of the study area. The CIE Lab color space and Otsu segmentation algorithm were used to detect the change of visible light DOM. Then, based on the theory of probability and statistics, an adaptive threshold determination method of 3-D change detection is put forward for the differential DEM in old and new times, and the high-risk area of geological disasters is extracted under the high-confidence confidence domain. Finally, the 3-D variation of earthwork is estimated by using the discretization integral method in high-risk areas. The experimental results show the feasibility, effectiveness and practicability of the proposed method. This paper not only sublimates the 2-D change of conventional remote sensing to 3-D space, but also quantitatively estimates the 3-D topography. It can be applied to the dynamic monitoring and assessment of geological disasters.