论文部分内容阅读
为了分析汶川县、茂县和理县地震重灾区的灾情状况,对震前2007年4月19日和震后2008年5月23日美国Landsat7ETM遥感图像进行辐射纠正、几何纠正、图像配准,然后使用Otsu最大类间方差阈值法二值化,分别通过二值图像逻辑异或处理、图像差值法自动检测地震造成的地表地物变化。结果表明,地震造成的地表地物变化主要沿岷江、杂谷脑河等河流两岸以及陡峭的沟谷地形地区,分布广且密集,其中杂谷脑河及流入其的溪谷地区的地表地物变化较大,地表地物变化的主要原因是地震引发了陡峭山体的大量崩塌、滑坡和坡面泥石流,灾情非常严重。遥感图像变化检测可有效检测出较大山体形态完整的块体滑坡的规模与形态。
In order to analyze the disaster situation of Wenchuan County, Maoxian County and Lixian County, the images of Landsat 7 ETM remote sensing images before the earthquake on April 19, 2007 and after the earthquake on May 23, 2008 were radically corrected, geometrically corrected and image registered. Then, Otsu maximum variance between the two classes using the threshold of binary method, respectively, through the binary image logical XOR processing, image difference method automatically detects changes in surface features caused by earthquakes. The results show that the change of surface features caused by the earthquake is mainly distributed along the banks of the rivers such as the Minjiang River and the Zagunao River and steep valley topography. The distribution of the surface features of the Zagunao River and the valleys flowing into it Larger, the main reason for the change of surface features is the earthquake caused a large number of landslides, landslides and slope debris flow, the disaster is very serious. The remote sensing image change detection can effectively detect the size and shape of the block landslide in the form of larger mountains.