Full-polarization SAR image ship detection method based on polarization characteristic driving

The invention discloses a full-polarization SAR image ship detection method based on polarization characteristic driving, and the method comprises the steps: S1, carrying out the preprocessing of a full-polarization SAR image, and obtaining the polarization characteristic data of the full-polarization SAR image; s2, constructing a convolutional neural network driven by polarization characteristics, inputting polarization characteristic data into a backbone network and a dense structure to extract polarization characteristics contained in the polarization characteristic data, and performing multi-scale fusion in a characteristic pyramid, the backbone network and the dense structure being used for context characteristic reuse in the network; and S3, according to preset training parameters and a loss function, training the convolutional neural network driven by the polarization characteristics to form a fully-polarized SAR image target detection network, and executing fully-polarized SAR image ship detection and