ISBN: 978-981-14-4787-7 DOI: 10.18178/wcse.2020.06.033
Ship Detection Method in Complex Sea Background
Abstract— in view of the influence of sea clutter, scale change of ship and blurred picture caused by environmental factors, this paper proposes a new method of ship detection based on saliency analysis and improved HOG. Firstly, the saliency map is obtained by two-dimensional Gabor multi-directional filtering, and the confidence interval of the multi-directional saliency map is calculated to segment the image, and the segmented image is fused to obtain the final regional proposal. Then, according to the influence of ship scale change and blurry, a feature extraction algorithm based on the combination of Gaussian pyramid and Hog is proposed. Compared with the traditional Hog features, this method has better robustness. Finally, the trained SVM model is used to block by block detection the regional proposal. The experimental results show that the detection accuracy of the algorithm is 93.91% in many scenes on the sea, which can effectively achieve the detection of ships on the sea
Index Terms— ship detection, saliency analysis,SVM classifier
Cuifang Zhao, Mengfan Che
Zhejiang Normal University, CHINA
Cite:Cuifang Zhao, Mengfan Che , " Ship Detection Method in Complex Sea Background " Proceedings of 2020 the 10th International Workshop on Computer Science and Engineering (WCSE 2020), pp. 214-218, Shanghai, China, 19-21 June, 2020.