多特征区域的细粒度船舶图像目标识别方法

doi:10.3778/j.issn.1002-8331.2010-0190

摘要/Abstract

摘要： 为解决单一特征细粒度船舶图像识别率低的问题，提出一种循环注意卷积神经网络（recurrent attention convolutional neural network，RA-CNN）与多特征区域融合的船舶目标识别方法。该方法通过在VGG-19网络中引入尺度依赖池化（scale-dependent pooling，SDP）算法解决小目标过度池化的问题，提升了小型船舶的识别性能；注意建议网络（attention proposal network，APN）加入联合聚类（joint clustering）算法，生成多个独立的特征区域，使整个模型充分利用全局信息，提高了船舶识别精度；同时设计特征区域优化方法降低多个特征区域的重叠率，解决了过拟合问题；通过定义新的损失函数来交叉训练VGG-19和APN，加快了收敛速度。利用公开的光电船舶数据集对该方法进行测试实验，识别准确率最高可达90.2%，无论是识别率还是模型的鲁棒性较单特征都有了很大的提升。

关键词: 船舶识别, 细粒度图像, 多特征区域, 循环注意卷积神经网络（RA-CNN）

Abstract: In order to solve the problem of low recognition accuracy in fine-grained ship images with a single feature, a ship target recognition method based on the fusion of recurrent attention convolutional neural network（RA-CNN） and multi-feature regions is proposed. This method introduces the scale-dependent pooling（SDP） algorithm in the VGG-19 network to solve the problem of excessive pooling of small targets in the network, and improves the recognition performance of small ships. The attention proposal network（APN） introduces joint clustering algorithm to form multiple independent feature regions, so that the whole model can make full use of global information and improve the accuracy of ship recognition. At the same time, a feature region optimization method is designed to reduce the overlap rate of multiple feature region and solve the problem of overfitting. By defining a new loss function to crosstrain the VGG-19 and APN, the convergence is accelerated. The proposed method is tested by using the open optoelectronic ship database, the recognition accuracy is up to 90.2%. Both the recognition rate and the robustness of the model are greatly improved compared with the single feature.

Key words: ship recognition, fine-grained image, multiple feature regions, recurrent attention convolutional neural network（RA-CNN）

徐志京, 孙久武, 霍煜豪. 多特征区域的细粒度船舶图像目标识别方法[J]. 计算机工程与应用, 2022, 58(10): 224-230.

XU Zhijing, SUN Jiuwu, HUO Yuhao. Fine-Grained Ship Image Target Recognition Method Based on Multiple Feature Regions[J]. Computer Engineering and Applications, 2022, 58(10): 224-230.

参考文献

[1] 张雷，甘春生，胡宇.高分辨率光学遥感影像舰船检测算法研究[J].计算机工程与应用，2017，53（9）：184-189.
ZHANG L，GAN C S，HU Y.Ship detection algorithm research on high resolution optical remote sensing image[J].Computer Engineering and Applications，2017，53（9）：184-189.
[2] 张仲瑜，焦淑红.多特征融合的红外舰船目标检测方法[J].红外与激光工程，2015，44（S1）：29-34.
ZHANG Z Y，JIAO S H.Infrared ship target detection method based on multiple feature fusion[J].Infrared and Laser Engineering，2015，44（S1）：29-34.
[3] 刘许，宋阳.多特征融合的舰船目标分类研究[J].舰船科学技术，2016，38（14）：88-90.
LIU X，SONG Y.Classification of ship based on multifeature fusion[J].Ship Science and Technology，2016，38（14）：88-90.
[4] 赵亮，王晓峰，袁逸涛.基于深度卷积神经网络的船舶识别方法研究[J].舰船科学技术，2016，38（15）：119-123.
ZHAO L，WANG X F，YUAN Y T.Research on ship recognition method based on deep convolutional neural network[J].Ship Science and Technology，2016，38（15）：119-123.
[5] 赵春晖，周瑶.基于改进Faster R-CNN算法的舰船目标检测与识别[J].沈阳大学学报（自然科学版），2018，30（5）：366-371.
ZHAO C H，ZHOU Y.Ship target detection and recognition based on improved Faster R-CNN algorithm[J].Journal of Shenyang University（Natural Science），2018，30（5）：366-371.
[6] CHO J H，PARK C G.Multiple feature aggregation using convolutional neural networks for SAR image-based automatic target recognition[J].IEEE Geoscience and Remote Sensing Letters，2018，15（12）：1882-1886.
[7] 乔丹，刘刚，杨执钧，等.基于迁移学习的船舶目标识别[J].计算机应用研究，2020，37（S1）：324-325.
QIAO D，LIU G，YANG Z J，et al.Ship target recognition based on transfer learning[J].Application Research of Computers，2020，37（S1）：324-325.
[8] ZHANG X，XIONG H，ZHOU W，et al.Picking deep filter responses for fine-grained image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition，2016：1134-1142.
[9] 熊昌镇，蒋杰.多尺度区域特征的细粒度分类算法研究[J].郑州大学学报（理学版），2019，51（3）：55-60.
XIONG C Z，JIANG J.Research on fine-grained classification algorithm of multi-scale regional features[J].Journal of Zhengzhou University（Natural Science Edition），2019，51（3）：55-60.
[10] ZHANG N，DONAHUE J，GIRSHICK R，et al.Part-based R-CNNs for fine-grained category detection[C]//13th European Conference on Computer Vision.Cham：Springer，2014：834-849.
[11] 胡志伟，杨华，黄济民，等.基于注意力残差机制的细粒度番茄病害识别[J].华南农业大学学报，2019，40（6）：124-132.
HU Z W，YANG H，HUANG J M，et al.Fine-grained tomato disease recognition based on attention residual mechanism[J].Journal of South China Agricultural University，2019，40（6）：124-132.
[12] FU J，ZHENG H，MEI T.Look closer to see better：recurrent attention convolutional neural network for fine-grained image recognition[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition，2017.
[13] 霍煜豪，徐志京.基于改进RA-CNN的舰船光电目标识别方法[J].上海海事大学学报，2019，40（3）：38-43.
HUO Y H，XU Z J.Photoelectric ship target identification method based on improved RA-CNN[J].Journal of Shanghai Maritime University，2019，40（3）：38-43.
[14] RUSSAKOVSKY O，DENG J，SU H，et al.Imagenet large scale visual recognition challenge[J].International Journal of Computer Vision，2015，115（3）：211-252.
[15] YANG F，CHOI W，LIN Y.Exploit all the layers：fast and accurate CNN object detector with scale dependent pooling and cascaded rejection classifiers[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition，2016：2129-2137.
[16] DU C，GAO S.Image segmentation based multi-focus image fusion through multi-scale convolutional neural network[J].IEEE Access，2017，5：15750-15761.
[17] 杨俊闯，赵超.K-Means聚类算法研究综述[J].计算机工程与应用，2019，55（23）：7-14.
YANG J C，ZHAO C.Survey on K-Means clustering algorithm[J].Computer Engineering and Applications，2019，55（23）：7-14.
[18] 蔡晓妍，戴冠中，杨黎斌.谱聚类算法综述[J].计算机科学，2008，35（7）：14-18.
CAI X Y，DAI G Z，YANG L B.Survey on spectral clustering algorithms[J].Computer Science，2008，35（7）：14-18.
[19] 董旭，魏振军.一种加权欧氏距离聚类方法[J].信息工程大学学报，2005，6（1）：27-29.
DONG X，WEI Z J.A clustering method of Euclid distance with weights[J].Journal of Information Engineering University，2005，6（1）：27-29.