Author(s)

Lianping Shan, Pan Jiang, Yihai Liu

Corresponding Author

Lianping Shan

ABSTRACT

In recent years, convolutional neural networks, CNN have become more and more excellent in the fields of image classification and object detection. The research on the application of deep learning in sea battlefield target recognition is more and more abundant. This paper first summarizes the theory and development process of the commonly used deep learning techniques in the target image recognition system. Then, compares traditional recognition technology and deep learning technology, two-stage model based on regional proposal and one-stage model based on regression. The status quo of deep learning technology in the sea battlefield target recognition is reviewed. Finally, we prospect the possible direction of future recognition technology of sea battlefield targets.

KEYWORDS

Convolutional neural network, deep learning, image recognition