Application of Image Data Analytics for Immediate Disaster Response

Abstract

This study identifies a novel source of data, i.e. images from smart urban infrastructures, that would be helpful in effective disaster management decision-making. For this purpose, we collected images from disaster-hit environments of Central Mexico (2017 earthquake). Also, this study utilizes deep learning convolutional neural network to analyze this novel dataset and evaluates the model effectiveness and technical viability during crisis scenarios. TensorFlow was utilized for the image classification task. The findings have important significance for effective disaster response.