Certain investigations on the application of neural networks for river water level prediction from satellite images

Abstract

The area of research undertaken in this work is to predict the water level in the river by using the satellite images In order to predict the water level artificial neural networks are used Yet using remotely sensed images for prediction of river water level involves a lot of preprocessing To begin with remote sensing makes it possible to collect data on newlinedangerous or inaccessible areas Remote sensing provides real time updates and it does not require active human assistance It can detect things that are not normally present in the visible spectrum for example detecting land forms underneath the surface of the ground or ocean An easy coverage of large areas of land as measurements can be made from a long distance and it is considered as one of the benefits of remote sensing In a few cases periodic observation of a target daily or even multiple times per day is possible only with satellite instruments The Artificial Neural Networks are information processing newlinemodels that are developed by taking inspiration from the working principles of the human brain The most essential property of ANN models is its ability to learn from sample sets newline