Flood Detection and Control Using Deep Convolutional Encoder-decoder Architecture
Husnu Baris Baydargil, Serkan Serdaroglu, Jang‐Sik Park, Kwang-Hee Park, Hyun-Suk Shin · 2018
In most major cities, there are certain areas with insufficient sewage systems that are incapable of moving drainage water efficiently, especially in heavy rain scenarios. Using CCTV (Closed-circuit television) cameras in order to detect such water pooling requires human focus and attention, however, immediate action has to be taken to minimize the environmental impact of such incidents. Detecting flooding not only might help save people's and city's money and time, but also human lives as well. In this paper, we propose a deep learning approach to detect flooding. Using the method of image segmentation and state-of-the-art architecture, system is capable of detecting flooding early on.