This paper presents a advance approach for ship detection in satellite imagery utilizing a modified DeepLabV3+ architecture, specifically designed to overcome the challenges inherent in such data. The proposed model features an enhanced feature extraction process and a refined atrous spatial pyramid pooling (ASPP) module, which together improve the detection of ships across various sizes and shapes. Comprehensive experiments on publicly available satellite datasets reveal that the modified DeepLabV3+ significantly outperforms existing state-of-the-art methods, achieving an accuracy of 98%. These findings demonstrate the model's robust ability to identify and localize ships in complex maritime settings, offering promising potential for improved maritime situational awareness and operational efficiency.
Enhancing Ship Detection on Satellite Images with Modified DeepLabV3+
04.09.2024
2666059 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Low-power Ship Detection in Satellite Images Using Neuromorphic Hardware
ArXiv | 2024
|Ship Detection in Optical Satellite Images Based on Sparse Representation
Springer Verlag | 2017
|