Design of Automatic Disaster Detection Algorithm for Remote Sensing Images Based on Deep Learning

In recent years, the emergence of DL (Deep Learning) technology has revolutionized the field of computer vision, surpassing the limitations of traditional approaches. This advancement has presented a novel solution for RS (Remote Sensing) image target detection and recognition challenges. Consequently, this article explores the utilization of CNN (Convolutional Neural Networks) as a cutting-edge technique for designing an automatic disaster area detection algorithm tailored for RS images. Initially, the article delves into the fundamental principles underlying DL algorithms and their applicability in identifying disaster-affected regions within RS imagery. Following this introduction, an innovative algorithm is devised that leverages DL for automatic disaster detection in RS images. This algorithm employs CNN to extract spatial features and capture sequential information pertaining to disaster areas. By integrating CNN within the DL framework, the algorithm maximizes the utilization of spatial and temporal data inherent in RS images, thereby enhancing the precision of disaster detection. To rigorously validate the efficacy of the proposed algorithm, several publicly available datasets are utilized for experimental purposes. The results obtained demonstrate that our algorithm outperforms existing ML (Machine Learning) and DL methodologies in terms of accuracy, recall, and F1 score. This research lays the groundwork for future explorations into the automated detection and recognition of disaster areas using CNN-based RS images.