YOLO-RS: An Efficient YOLO-Based Approach for Remote Sensing Object Detection

This research introduces “YOLO-RS,” an efficient object detection system based on YOLO tailored for remote sensing applications. YOLO-RS excels in real-time object detection in satellite and aerial imagery, optimizing performance while maintaining high accuracy. This study evaluates YOLO-RS on diverse remote sensing datasets, showcasing its precision and recall. The framework's speed and adaptability make it a valuable tool for tasks such as land cover mapping and disaster response in remote sensing scenarios. In this analysis, we have chosen the SENTINEL-2 dataset. With the help of the SENTINEL-2 dataset, we have evaluated the performance and accuracy of YOLOv3 and analyzed its strengths and weaknesses. Using the results obtained from the implementations, we determined and optimized the detection of building and trees in satellite imagery. The computed results indicate an average accuracy of 85.6%, precision of 98.9%, and recall of 86.6%.