Ship target detection method based on improved DCN

The invention discloses a ship target detection method based on an improved DCN, and the method comprises the steps: selecting a remote sensing image data set with a marking file containing direction information as an input image, and carrying out the random overturning and filling of the input image; preprocessing the input image, and dividing the input image into a training set, a verification set and a test set; inputting the training set into the improved DCN model for training; inputting the test set into the trained improved DCN model, detecting the remote sensing image to obtain a directed bounding box, and detecting a ship target; according to the invention, the DCN is adopted to obtain the offset of the point set, the APAM module is adopted to screen positive samples, the SPP module is introduced to realize the fusion between local features and global features, the correlation of classification and positioning is further enhanced, the acquisition of high-quality samples is ensured, and the high-preci