目前舰船识别技术相对落后，大部分识别算法还是基于传统的机器学习理论。舰船识别受舰船背景、光照、遮挡等因素的影响较大，识别正确率较低，无法满足现实需求；随着深度卷积神经网络识别率正确率的不断提高，一些复杂的分类任务都得到了较好的解决。本文将深度卷积神经网络AlexNet迁移到舰船识别中，对原网络顶层进行改进，微调底层特征，采用数据扩充技术构建的舰船数据集训练调试模型，得到了有效的舰船识别模型，识别正确率达到91.08%。
The technology of ship recognition has relatively dropped behind. Most of the recognition algorithms are based on traditional machine learning theory. However, due to the influence of environment, illumination and occlusion, accuracy of recognition ships is low and can't meet the actual needs. With the improvement of the accuracy of recognition rate by convolutional neural networks in depth learning, some complicated tasks of recognition have been solved well. In this paper, the convolutional neural network is transferred to the task of recognition ship. The top layer of the original model is improved and the characteristics of the bottom layers are finely tuned. Besides, the model trained with a limited data set of ships can obtain a recognition accuracy of 91.08%.
2018,40(10): 118-121 收稿日期：2018-01-25
DOI：10.3404/j.issn.1672-7649.2018.10.023
分类号：TP391.4
基金项目：海军装备部“十二五”规划资助项目（435515908）
作者简介：邢世宏(1989-),男,博士研究生,研究方向为计算机视觉、图像处理、深度学习
参考文献：
[1] 王彦情, 马雷, 田原. 光学遥感图像舰船目标检测与识别综述[J]. 自动化学报, 2011, 37(9):1029-1039.
[2] 周珍娟, 韩金华. 舰船遥感图像的目标识别研究[J]. 舰船科学技术, 2014, 36(12):86-90.
[3] 赵志刚. 舰船自动识别系统中的网络安全技术研究[J]. 舰船科学技术, 2016, 38(5A):121-123.
[4] 赵矿所, 林钧清. 神经网络在舰船噪声识别中的应用[J]. 舰船科学技术, 1995, 17(6):32-36.
[5] KRIZHEVSKY A, SUTSKEVER I, HINTON G E. Imagenet classification with deep convolutional neural networks[J]. Communications of the Acm, 2012, 60(2):2012.
[6] LAB S V. ILSVRC[EB/OL]. http://image-net.org/challenges/LSVRC/.
[7] Stanford Vision Lab S U, Princeton University ImageNet[EB/OL]. http://www.image-net.org/.
[8] HE K, ZHANG X, REN S, et al. Deep Residual Learning for Image Recognition[C]. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2015: 770-778.
[9] ZENG X, HUANG H, QI C. Expanding training data for facial image super-resolution[J]. IEEE Transactions on Cybernetics, 2017, PP(99):1-14.
[10] WU Y N. Data augmentation[J]. Computer Vision, 2014:165-166.
[11] FAWZI A, SAMULOWITZ H, TURAGA D, et al. Adaptive data augmentation for image classification[C]//IEEE International Conference on Image Processing, 2016:3688-3692.
[12] JIA Y A S, Evan and Donahue, Caffe[EB/OL]. http://caffe.berkeleyvision.org/.
[13] ZEILER M D, FERGUS R. Visualizing and understanding convolutional networks[M]. Springer International Publishing, 2014:818-833.
