Application of Image Analytics for Disaster Response in Smart Cities

Abstract

Post-disaster, city planners need to effectively plan response activities and assign rescue teams to specific disaster zones quickly. We address the problem of lack of accurate information of the disaster zones and existence of human survivors in debris using image analytics from smart city data. Innovative usage of smart city infrastructure is proposed as a potential solution to this issue. We collected images from earthquake-hit smart urban environments and implemented a CNN model for classification of these images to identify human body parts out of the debris. TensorFlow backend (using Keras) was utilized for this classification. We were able to achieve 83.2% accuracy from our model. The novel application of image data from smart city infrastructure and the resultant findings from our model has significant implications for effective disaster response operations, especially in smart cities.