基于RADARSAT-1影像的洪涝评估方法

A Flood and Water-logging Assessment Method Based on RADARSAT-1 Image

RADARSAT-1雷达卫星因其全天时、全天候、多种地面分辨率的下垫面信息采集能力迅速在防洪减灾领域得到应用。本文基于淮河流域2007年梅雨期的RADARSAT-1影像,利用遥感影像处理软件ENVI的改进型弗罗斯特滤波器对其进行滤波运算,消除其噪音斑点。然后基于ENVI的决策树分类器,对下垫面进行水体和非水体的划分。把RADARSAT-1影像上的水体区域减去背景SPOT5影像的水体后,获得了本次洪涝的区域范围。最后根据洪涝范围内的高精度数字地形图、人口分布和耕地分布等基础数据,估算出主要行蓄洪区的淹没耕地面积和影响人口等灾情,从而为防洪减灾工作提供了信息支持。

RADARSAT-1 has been applied in flood control for its characteristics of all-weather, all-day and multiresolution land surface image obtain capacity. Based on the RADARSAT-1 data obtained from the Huaihe River Basin during the plum rains period of 2007, the Enhance Frost Filter package of ENVI aremote sensing image process software, was used to reduce its noise speckles. After filtering, the decision tree package of ENVI was used to classify the land surface into water-body and others. After detracting the water-body classified with SPOT5, the background image, from that of RADARSAT-1, the flood and water-logging area was well-marked. Eventually, based on the fundamental data of the flood and water-logging area, such as high-precision digital line graphics, population and its distribution, tillage and its distribution and etc, the submersed tillage area and affected population were estimated. The results show that the model can provide information support for flood control and disaster reducing.