Abstract: Flood are the most common type of natural disaster, and causes thousands of casualities every year in the world. Flood causes approximately 30% loss in total disaster. In this article, We have shown to save lives, where a Novel Image Classification techniques for Satellite images using Modified CNN – Capsnet to improve the accuracy of the output. It provides faster information supply to the user, collect accurate flood event and in cost – effective manner, The proposed work helps to classify the inundates areas efficiently in order to estimate and plain relief work rapidly.

Keywords: Deep learning, Capsule Network, Image classification, Object detection, Satellite images.

| DOI: 10.17148/IJIREEICE.2020.8813