A remote sensing ship recognition method based on dynamic probability generative model

作者：

Highlights：

• An initial contour extraction based on visual saliency prior shape is introduced.

• Based on entropy and local neighborhood information, CV model is improved.

• Based on rough set theory, common discernibility degree is used to select features.

• Probability generative model with neighbor nodes’ classes is used to recognize ships.

摘要

•An initial contour extraction based on visual saliency prior shape is introduced.•Based on entropy and local neighborhood information, CV model is improved.•Based on rough set theory, common discernibility degree is used to select features.•Probability generative model with neighbor nodes’ classes is used to recognize ships.

论文关键词：Ship recognition,Saliency,Image segmentation,Entropy,ε-Local neighborhood information,Probability generative model

论文评审过程：Available online 13 April 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.03.033