Regional-Scale Object Feature Recognition in High-Resolution Remote Sensing Images Based on YOLOv8

Accurate object detection in regional-scale remote sensing imagery is challenged by significant scale variation, dense small targets, and complex background interference. While mainstream detection frameworks have achieved competitive results, their performance often declines when directly applied to high-resolution aerial scenes. To address this issue, this paper presents an enhanced YOLOv8-based detection framework and conducts a systematic comparative evaluation against representative state-of-the-art methods. A lightweight attention-guided multi-scale feature refinement module is introduced to strengthen small object representation, and an optimized feature aggregation strategy is adopted to improve localization precision without increasing computational complexity. Experiments are performed on two widely used remote sensing datasets, NWPU VHR-10 and DIOR. The proposed method achieves consistent performance improvements over competitive mainstream detectors, with approximately 1.98% gain in mAP while preserving real-time efficiency. Ablation studies further confirm the effectiveness of the proposed refinement mechanism. The results demonstrate that the improved YOLOv8 framework provides a robust and efficient solution for regional-scale remote sensing object detection.