Introduction:
The AI Theory class has provided us with comprehensive foundation in artificial intelligence, covering both fundamental principles and advanced methods. Through topics such as clustering, neural networks, evolutionary computing, and decision-making models, the course aims to equip students with the theoretical knowledge and practical insights needed to apply AI across various fields, including disaster management and urban planning.
Abstract:
SeisNAV aims to be an innovative disaster management system designed to assist NGOs and emergency responders by providing crucial real-time information during crises. It leverages AI to detect building collapses in disaster-stricken areas and updates maps to identify blocked roads, enabling efficient rerouting from affected locations to critical facilities like hospitals. By addressing the immediate challenges of navigation and resource allocation, SeisNAV lays the foundation for more effective disaster response. As a next step, we propose integrating drones into the workflow to enhance data collection and further support stakeholders in making informed, life-saving decisions.
PROBLEM TO SOLVE
TARGET
CURRENT SOLUTION
DEVELOPING THE IDEA FURTHER
References:
Satellite images dataset: https://www.maxar.com/
Paper:
Convolutional neural networks for object detection in aerial imagery for disaster response and recovery – Yalong Pia, Nipun D. Nathb, Amir H. Behzadana. https://www.sciencedirect.com/science/article/abs/pii/S1474034619305828
Disaster Management Redefined: Integrating SVM‑AE Techniques with Remote Sensing and Meteorological Data – L. Priyadharshini, Jyoti A. Dhanke, R. N. Patil, B. Swapna, Kapula Kalyani, Maganti Syamala, Shanmugavel Deivasigamani.
https://www.researchgate.net/publication/383658602_Disaster_Management_Redefined_Integrating_SVM-AE_Techniques_with_Remote_Sensing_and_Meteorological_Data
Airborne LiDAR point cloud classification using PointNet++network with full neighborhood features – Xingzhong Nong, Wenfeng Bai
https://www.researchgate.net/publication/368425147_Airborne_LiDAR_point_cloud_classification_using_PointNet_network_with_full_neighborhood_features
