Main Article Content
Rainfall is a natural phenomenon and its observations are frequently essential for flood warning and its monitoring. The conventional techniques which are used for rain measurement is rain gauge, radar and satellite, precipitation estimation with each technique has its own asset with restriction and the methods are supportive to some extent, providing information at different spatial and terrestrial scales. In this article, a new approach is developed to estimate rainfall from droplet images data using proposed signal processing algorithm. This proposed method uses multi-sensor image data captured by two orthogonal placed cameras. After selection of appropriate data image, it is enhanced for noise free and smooth edges then. Fusion of multi sensor image data gets more informative image followed by segmentation for rainfall parameter analysis. Finally, drop size distribution, rainfall rate and direction of precipitation with its intensity are computed using the proposed algorithm. The real time image data in the form of images / videos were collected at Indian Meterology Centre, Coloba Mumbai, during the day time from 24th July to 6th August 2016 and experiment results of the method are compared with rain gauge observation techniques at the same place. We found that mathematical computations are only involved in this method and most convenient alternative by portable (cellular) device over conventional tedious and laborious methods. It would be additional helpful to the disaster administrators the city, water dam administrators, traveler, piscatorial and farmers to know rainy condition in every side in advance.