Flood Water Depth Classification Using Convolutional Neural Networks

摘要

Floods cause havoc in many regions of India every year during the monsoon season. However, commuters really do not really think of the after effects caused by the floods. Commuting during water-logging is risky because the depth of logged water cannot be determined. This can lead to cars and bikes getting stuck inside the flooded zone. Thus travelling through such areas can lead to loss of life. Our paper focuses on the classification of floodwater depth. We propose a deep-learning model that classifies the depth of the flood water level into different categories according to the depth. In this paper, we have used the Convolutional Neural Network (CNN) to classify the images given as an input by the user/commuter. For better user accessibility we propose to integrate the model with a cross-platform app that would be convenient for the commuters; along with an in-app alert system, which notifies other users about the location where the water level is above threshold value.