Ship Image Denoising Algorithm Research Based on Deep Learning

During the voyage, the ship is vulnerable to the influence of the external environment, and the image quality of the obtained ship needs to be improved, which affects the ship identification to a certain extent. Based on the deep learning DnCNN algorithm, an improved denoising algorithm is proposed. In order to improve the training range of neurons, Leakly ReLU function is used instead of ReLU activation function in network training to extract and learn the features of the input image, so as to improve the fitting ability of the network. The training network of DnCNN algorithm is simplified and the residual network is added to make the network training faster and more stable. The experimental results show that the improved algorithm has a better denoising effect on ship image.