Ship Recognition Method based on Visual Knowledge Base
Sui Ting-ting, Xi Zhuang-Ying-Zi, Chen Bang-Chao, Hua Qiao, Xie Ping · 2018
In order to ensure the navigational safety, a ship recognition model based on visual knowledge base (SRVKB) is presented. In SRVKB, CNN model is combined with Itti model and multi-feature extraction method. Moreover, SRVKB can yield a visual knowledge base of ship, which simulates the process from visual attention to object recognition in human visual system. Firstly, carry out de-noising process and contour extraction on the object area of the image. Secondly, by adding three different channels to CNN model, multi-features can be obtained. Finally, visual knowledge base of ship yielded via SRVKB can be used as a guide in ship recognition. The experiment results show that the proposed method can not only locate the ships' positions. Meanwhile, compared with single feature method, SRVKB with multi-features can better the recognition effective.