Multi-angle SAR to Sea Recognition Algorithm Based on Parallel Deep Learning

Synthetic aperture radar (SAR) is widely used in maritime target detection due to its all-weather and all-day detection capability. Due to the sensitivity of SAR images to the imaging azimuth, the SAR images of the same target in different azimuths have certain differences, and the SAR image sequences with multiple azimuths contain richer ship target information. Therefore, this paper proposes a ship target recognition method based on the combination of optimal mapping feature extraction and parallel deep learning. The method has better recognition accuracy than other target recognition on most of the datasets, and has certain robustness, strong fault tolerance to open environment, data loss, corruption and high recognition accuracy in real scenarios.