Disaster management using deep learning on social media

Abstract

ABSTRACT The main goal of this proposed work is to provide solutions for disaster management using deep learning algorithms on social media images. The MNIST dataset was used to initially build the deep learning models. The images were trained using LeNet5, VGG13, VGG 16 and LSTM deep learning models. Later a dataset containing 3460 images were taken from social media. The labels earthquake, wildfire and floods were used to achieve classification results. The images were trained and validated using LSTM, VGG13 and VGG16. The performance of the algorithms is compared and the disaster response technique is generated based on the image classification and disaster management strategies are provided based on classification.