Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

FloodLense: A Framework for ChatGPT-based Real-time Flood Detection

Pranath Reddy Kumbam, Kshitij Maruti Vejre

TL;DR

FloodLense presents an integrated framework that tackles real-time flood detection by combining deep learning architectures (UNet, RDN, ViT) with a ChatGPT-powered conversational interface. The approach processes aerial and satellite imagery, leveraging Sentinel Hub data and a plug-in-like workflow to deliver timely flood mappings and natural language-based queries. Key contributions include demonstrable improvements in flood delineation, an end-to-end architecture linking image analysis with user-friendly interaction, and a thorough evaluation across FloodNet and Sentinel datasets, highlighting model strengths, limitations, and practical deployment considerations. The work holds significant potential for real-time disaster management and accessible environmental monitoring, enabling non-expert users to obtain actionable flood information via conversational AI.

Abstract

This study addresses the vital issue of real-time flood detection and management. It innovatively combines advanced deep learning models with Large language models (LLM), enhancing flood monitoring and response capabilities. This approach addresses the limitations of current methods by offering a more accurate, versatile, user-friendly and accessible solution. The integration of UNet, RDN, and ViT models with natural language processing significantly improves flood area detection in diverse environments, including using aerial and satellite imagery. The experimental evaluation demonstrates the models' efficacy in accurately identifying and mapping flood zones, showcasing the project's potential in transforming environmental monitoring and disaster management fields.

FloodLense: A Framework for ChatGPT-based Real-time Flood Detection

TL;DR

FloodLense presents an integrated framework that tackles real-time flood detection by combining deep learning architectures (UNet, RDN, ViT) with a ChatGPT-powered conversational interface. The approach processes aerial and satellite imagery, leveraging Sentinel Hub data and a plug-in-like workflow to deliver timely flood mappings and natural language-based queries. Key contributions include demonstrable improvements in flood delineation, an end-to-end architecture linking image analysis with user-friendly interaction, and a thorough evaluation across FloodNet and Sentinel datasets, highlighting model strengths, limitations, and practical deployment considerations. The work holds significant potential for real-time disaster management and accessible environmental monitoring, enabling non-expert users to obtain actionable flood information via conversational AI.

Abstract

This study addresses the vital issue of real-time flood detection and management. It innovatively combines advanced deep learning models with Large language models (LLM), enhancing flood monitoring and response capabilities. This approach addresses the limitations of current methods by offering a more accurate, versatile, user-friendly and accessible solution. The integration of UNet, RDN, and ViT models with natural language processing significantly improves flood area detection in diverse environments, including using aerial and satellite imagery. The experimental evaluation demonstrates the models' efficacy in accurately identifying and mapping flood zones, showcasing the project's potential in transforming environmental monitoring and disaster management fields.
Paper Structure (16 sections, 7 figures, 7 tables)

This paper contains 16 sections, 7 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Societal Applications
  3. Existing Methods
  4. Technical Contributions and Project Novelty
  5. Limitations of Current Approaches
  6. Problem definition
  7. Proposed solution
  8. Psuedo-Code for Satellite Image Processing and Retrieval
  9. Psuedo-Code to Simulate ChatGPT Plugin
  10. Evaluation
  11. Evaluation Data and Experimental Configuration
  12. Evaluation Metric
  13. Performance Analysis and Results
  14. Conclusion
  15. Appendix
  16. ...and 1 more sections

Figures (7)

  • Figure 1: ChatGPT plugin Simulation
  • Figure 2: Original vs Highlighted Flood Detected Image
  • Figure 3: Sample Output Showing Flooded Buildings - UNet (Top), RDN (Bottom)
  • Figure 4: System Architecture
  • Figure 5: Sample Outputs from FloodNet Data. Positive values indicate water and flooded regions.
  • ...and 2 more figures