Abstract
Hurricanes or tropical cyclones are very destructive and powerful storms accompanied by strong winds of 119 km/hours or greater. They are also accompanied by rain, thunder, and lightning. This causes huge damage to human lives as well as property. Satellite images are being used for damage detection, but manual methods are prone to errors. Deep learning techniques are gaining popularity for automatic damage detection. With this motivation, a comparative analysis of two Convolutional Neural Networks that is 8-layer model and 16-layer model has been done to classify Hurricane Harvey satellite images into undamaged and damaged categories. It was found that the 16-layer model achieved the best results with an accuracy of 94% and recall of 93%.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call