Detection of the flood boundary in SAR image using texture

SAR is, often used to monitor flood event. After SAR image is acquired, it is critical to capture flood area for response, recovery and mitigation activities during and after I flood event. The first step is to detect boundary of flood to survey flood area. However, it is difficult to detect flood boundary in some area due to excessive rainfall. In this paper, we propose a method to detect flood boundary based on texture. Because the texture of water area is different to texture of nowater area, we can use the texture information to detect flood boundary. The key step of the proposed method is to compute ratio of angular second moment and entropy, which will sharpen the edge with slow variation. Application of the proposed method to Huai river flood of China in 2003 indicate that the proposed method can effectively detect flood boundary.