Real-Time Object Detection for Rover and Drone: A Comparative Study from R-CNN to YOLO

Abstract

Real-time object detection is a fundamental requirement for autonomous navigation and surveillance systems. This work presents a comparative study of R-CNN, DCNN, and YOLO-based object detection frameworks implemented on an ADAS-enabled rover and a drone surveillance platform. Performance is evaluated in terms of inference latency, detection stability, and visual clarity under real-time operating conditions. Experimental results demonstrate that YOLO provides superior real-time performance while maintaining reliable detection accuracy, making it highly suitable for time-critical autonomous applications.