An Improved YOLOv8-based Ship Detection Algorithm for High-resolution Satellite Images
JIANG Zhihao
WEI Qiang
Abstract：To address the challenges of multi-scale object detection,insufficient feature extraction for small targets,and inter-ference from complex backgrounds in high-resolution satellite images,this paper proposes an improved YOLOv8-based ship detec-tion algorithm.In the feature extraction network,an efficient channel attention fusion module(ECAF)based on attention mecha-nisms is introduced to enhance the network's ability to extract ship target features.Additionally,in the neck network,an improved bi-directional feature pyramid network(BiFPN)is employed to optimize feature fusion and improve the efficiency of multi-scale fea-ture transmission.Experimental results demonstrate that the improved algorithm achieves mean average precision(mAP50)of 92.1%and 93.7%on the DIOR CV and NWPU-RESISC45 datasets.Compared to the original YOLOv8 algorithm,it has improved by 4.7%and 3.8%respectively.These results effectively enhance the performance of ship target detection in high-resolution satel-lite images.
Keywords：high-resolution satellite imagesship detectionattention mechanismfeature fusion
Publication Date：2025-07-20
Online Publishing Date：2025-09-25(First online date of this platform, not the publication date of the document)
Pages：7( 41-47 )