A Comparative Analysis of Post-Disaster Analysis Using Image Processing Techniques

Abstract

Post-disaster recovery is a multifaceted system essential for rebuilding communities and infrastructure. Despite its importance, many limitations obstruct powerful recuperation, main to tremendous loss of life and monetary assets. This paper synthesizes varied approaches in the direction of sustainable restoration, highlighting the increasing reliance on technology for disaster management. Image processing strategies, pivotal in addressing these demanding situations, are reviewed across studies. Those strategies range from SLIC segmentation and Random forest classification to advanced deep learning models together with U-net and YOLOv8, machine learning algorithms like SVM, and image category methodologies along with bi-temporal analysis. Comparative evaluation reveals that those strategies presents promising consequences, with accuracies starting from 75% to over 94%. The paper gives a framework for understanding the role of various image processing strategies in improving disaster control strategies, emphasizing their implications for future studies and application.

Keywords

• disasters
• post-disasters
• recovery
• machine learning techniques
• classifications