Enhanced Thermal Object Tracking in Search and Rescue Missions Using Adaptive 3D Kalman Filtering

Search and rescue (SAR) operations often involve hazardous environments where rapid human detection is critical for survival. While thermal imaging drones have improved SAR capabilities, challenges persist in maintaining accurate tracking amid dynamic environmental conditions, occlusions, and changing perspectives. This paper presents a novel approach to thermal human tracking by implementing an adaptive 3D Kalman filtering system integrated with depth estimation techniques. Our methodology enhances tracking accuracy by extending traditional 2D Kalman filters to incorporate altitude changes and advanced occlusion handling. Experimental results demonstrate significant improvements in detection reliability and tracking persistence across challenging scenarios including darkness, smoke, and dense foliage. The system achieves 92% detection accuracy in low-visibility conditions and reduces false positive rates by 37% compared to conventional methods. This research addresses critical gaps in SAR technology while offering applications across disaster management, wildlife conservation, and industrial safety domains. The proposed framework enables more efficient SAR operations, potentially reducing response times and improving survival outcomes in disaster scenarios.