Flood Estimation from Water Discharge Using Regression Analysis

Abstract

Flood is the maximum flow which may be run off a catchment area during a period of exceptional rainfall. The flood estimate models predicts the risk reduction with minimization of the loss of human life, and decrease of the property impairment linked with Flood. In this paper the flood is estimated by calculating image processing parameters on water discharge from canal. The estimation of flood is analysed with ‘Logistic Regression’, a machine learning algorithm and compared with existing state-of-art methods. This is then used to train and test the model. The work shown in this paper is performed on the Khadakwasala canal in Pune.