Mapping Nighttime Flood from MODIS Observations Using Support Vector Machines

作者：Zhang Rui^*; Sun Donglian; Yu Yunyue; Goldberg Mitchell D

来源：Photogrammetric Engineering and Remote Sensing, 2012, 78(11): 1151-1161.

DOI：10.14358/PERS.78.11.1151

This work proposes a nighttime flood mapping method for Moderate Resolution Imaging Spectroradiometer (MODIS) data. Brightness temperatures at 3.9 mu m, and 11 mu m channels (BT 3.9, and BT 11, respectively) and differences of brightness temperatures between 3.9 Am and 4.0 mu m, and between 11 mu m and 12 mu m (BT 3.9-BT 4.0, and BT 11-BT 12, respectively) are used to identify nighttime water pixels by a support vector machines (SVM) classifier. Prominent flood locations are detected by a change detection process using a reference water-land map. To test the effectiveness of the proposed method, two flood cases caused by heavy rains were chosen as trial scenarios. The nighttime mapping results are validated with the flood maps, which are obtained from the visual interpretation based on the daytime flood identification results. The experimental results indicate that the proposed method is effective for the delineation of inundated areas with standing water during the nighttime.

• 出版日期2012-11