A Practical Framework for Rapid Earthquake Damage Estimation through Multi-Vehicle Coordination and Inter-Agency Collaboration: Integration of Genetic Algorithms and RGB Image Analysis

Rapid and accurate assessment of building damage following large-scale earthquakes is essential for effective emergency response. This study proposes a practical framework that integrates camera-equipped ground vehicles deployed by multiple agencies (Self-Defense Forces, police, fire departments, etc.) with a central command system to map damage conditions in real-time immediately after an earthquake. This system combines a generalized Normalized Chromatic Difference Index – an RGB-based damage detection technology that extends the simplicity of the Redness Index developed for vegetation analysis to enhance the versatility of color indices – with a Genetic Algorithm that optimizes the patrol routes of multiple vehicles. Color-based inference rapidly identifies collapsed buildings from aerial imagery by detecting spectral characteristics of exposed timber and debris. Meanwhile, the GA efficiently assigns routes to each vehicle, maximizing coverage and minimizing response time. A cloud-based architecture standardizes geotagged damage reports in JSON format for real-time sharing across response agencies. This paper presents the system design, implementation details, and evaluation protocol based on disaster scenario simulations targeting Japan's Noto Peninsula region. Cross-validation-based accuracy achieved an F1 score of 0.84, detecting 93% of collapsed buildings while keeping false alarm rates around 23%. Simultaneously, the cooperative vehicle routing strategy significantly improved survey efficiency, reducing total mission completion time by approximately 25% compared to the greedy algorithm baseline. Furthermore, the authors discuss practical challenges such as speed and resolution advantages over conventional satellite and aerial assessments, data privacy considerations, false detections, and the necessity for human verification of results. Overall, this study demonstrates a feasible multi-vehicle, multi-agency approach for rapid earthquake damage estimation aimed at accelerating life-saving operations and optimizing resource allocation.