Development of a Real-Time Object Detection and Tracking System for an Autonomous Grenade Launcher Using Frame Differencing and Threshold Optimization

Development of a Real-Time Object Detection and Tracking System for an Autonomous Grenade Launcher Using Frame Differencing and Threshold Optimization

This study developed a real-time object detection and tracking system for an autonomous grenade launcher using frame differencing and adaptive thresholding. The system was designed to efficiently track moving objects in dynamic environments, focusing on human movement recognition. To assess its effectiveness, experiments were conducted by varying threshold values and analyzing their impact on detection accuracy. The results confirmed that a threshold of 30 optimally detected human movements while minimizing noise. Object detection experiments included analyzing detection results, cumulative motion visualizations, and object separation after background removal. The system achieved Precision 66.7%, Recall 88.9%, and F1 Score 76.2%, demonstrating reliable performance under general conditions. A comparison with a standard performance classification table further validated its accuracy. These findings suggest that the proposed method can be optimized for real-world applications requiring precise and robust object tracking.