Citation

Gupta N, Rana KK. Webology 2021; 18(5): 274-292.

Copyright

(Copyright © 2021, University of Tehran, Iran, Publisher Info Sci Publisher)

DOI

unavailable

PMID

unavailable

Abstract

Convolution Neural Network (CNN) is shadow-resistant, capable of getting proper disaster characteristics, and, most importantly, capable of overcoming operator misdetection or misjudgement, all of which have an impact on the system's efficacy. Training and testing are the two phases of the neural network. By clipping and resizing aerial pictures acquired from Kaggle, pre- and post-disaster training data patches are created. Focusing on Hurricane's Pacific Ocean Training dataset, which contains 4387 train photos. All patches are trained in CNN to extract disaster regions occurred without delay. The technique will detect disasters with an accuracy of 80-90%. An off-grid communication system is created with the help of a scatternet based on Bluetooth. This will help in establishing communication between various affected groups and can help in spreading the information faster to the authorities and rescue teams. The app will create the network required for communication among groups and the authorities and the website will display all the collected data to the concerned authorities. It will help in increasing the connectivity between affected people at the disaster location. It will allow the relief management team to have an upper hand by enabling them to contact the people in the affected area.

Language: en