基于改进YOLOV3-Tiny的海面船舰目标快速检测

doi:10.19678/j.issn.1000-3428.0059305

摘要/Abstract

摘要： 为实现海面船舰目标的快速、准确检测，提出一种改进的船舰目标检测算法。在网络结构方面根据船舰目标的特点，对浅层信息进行强化重构以降低小目标的漏检率，同时引入改进的残差网络增加网络深度和降低网络参数计算量，并且采用金字塔网络进行多尺度特征融合，以兼顾图像中大小船舰目标的检测性能。在网络训练中利用迁移学习策略进行网络模型的训练，以克服船舰图像样本集有限的问题。在视频检测中利用帧间图像结构相似度进行选择性网络前向计算，以提高视频帧检测速率。实验结果表明，该算法海面船舰目标检测的准确率达到92.4%，较YOLOV3-Tiny提高7个百分点，召回率达到88.6%，且在CPU平台上船舰目标的检测速度达到12 frame/s。

关键词: 卷积神经网络, YOLO网络, 船舰目标检测, 迁移学习, 深度学习

Abstract: In order to achieve fast and accurate detection of surface ship targets, this paper proposes a ship target detection algorithm based on improved YOLOv3-Tiny.Firstly, in network structure, the features of shallow layers of the network is enhanced and reconstructed according to the characteristics of ship targets to reduce the miss detection rate of small targets, and the improved residual network is introduced to improve the depth of the network while reducing the calculation of network parameters.Moreover, the pyramid network is used for multi-scale feature fusion to balance the detection capability between large ship targets and small ship targets in images.Secondly, in the network training, transfer learning strategy is employed to train the designed network model to alleviate the limitation of known ship image samples.Finally, in video detection, a video frame selection method for forward computation of the network model based on structure similarity of inter frames is proposed to improve the detection frame rate.The experimental results show that the proposed algorithm has precision rate up to 92.4%, with an increase of 7% compared with YOLOV3-Tiny, recall rate up to 84%, and detection frame rate up to 12 frames/s on CPU platform.

Key words: Convolutional Neural Network(CNN), YOLO network, ship target detection, transfer learning, deep learning

中图分类号:

TP391.41

李庆忠, 徐相玉. 基于改进YOLOV3-Tiny的海面船舰目标快速检测[J]. 计算机工程, 2021, 47(10): 283-289,297.

LI Qingzhong, XU Xiangyu. Fast Target Detection of Surface Ship Based on Improved YOLOV3-Tiny[J]. Computer Engineering, 2021, 47(10): 283-289,297.

http://www.ecice06.com/CN/Y2021/V47/I10/283

图/表 12

20211016175336

20211016175341

20211016175345

20211016175349

20211016175353

20211016175357

20211016175400

20211016175404

20211016175408

20211016175412

20211016175416

20211016175419

参考文献

[1] WANG C B, ZHANG X Y, CONG L Z, et al.Research on intelligent collision avoidance decision-making of unmanned ship in unknown environments[J].Evolving Systems, 2019, 10(4):649-658.
[2] 黎经元, 厉小润, 赵辽英.基于边缘线分析与聚合通道特征的港口舰船检测[J].光学学报, 2019, 39(8):225-234. LI J Y, LI X R, ZHAO L Y.Port ship detection based on edge line analysis and aggregation channel characteristics[J].Acta Optica Sinica, 2019, 39(8):225-234.(in Chinese)
[3] 李浩谊, 马春庭.基于改进的Scharr算法的海上舰船图像边缘检测[J].舰船电子工程, 2019, 39(3):103-106. LI H Y, MA C T.Image edge detection of marine ships based on improved Scharr algorithm[J].Ship Electronic Engineering, 2019, 39(3):103-106.(in Chinese)
[4] ZHANG Y, LI Q Z, ZANG F G.Ship detection for visual maritime surveillance from non-stationary platforms[J].Ocean Engineering, 2017, 141(1):53-63.
[5] 丁鹏, 张叶, 贾平, 等.基于视觉显著性的海面舰船检测技术[J].电子学报, 2018, 46(1):127-134. DING P, ZHANG Y, JIA P.Detection technology of sea surface ships based on visual saliency[J].Chinese Journal of Electronics, 2018, 46(1):127-134.(in Chinese)
[6] SHI G M, SUO J D.Ship target detection based on visual attention[C]//Proceedings of International Conference on Signal Processing, Communications and Computing.Qingdao, China:[s.n.], 2018:1-4.
[7] SHAO Z F, WANG L G, WANG Z Y, et al.Saliency-aware convolution neural Network for ship detection in surveillance video[J].IEEE Transactions on Circuits and Systems for Video Technology, 2019, 30(3):1-15.
[8] ZHANG W, HE X J, LI W Y, et al.An integrated ship segmentation method based on discriminator and extractor[J].Image and Vision Computing, 2019, 93(1):1-12.
[9] WANG Y C, NING X Y, LENG B H, et al.Ship detection based on deep Learning[C]//Proceedings of International Conference on Mechatronics and Automation.Tianjin, China:[s.n.], 2019:275-279.
[10] 马啸, 邵利民, 金鑫, 等.改进的YOLO模型及其在舰船目标识别中的应用[J].电讯技术, 2019, 59(8):869-874. MA X, SHAO L M, JIN X, et al.Improved YOLO model and its application in ship target recognition[J].Telecommunications Technology, 2019, 59(8):869-874.(in Chinese)
[11] 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, 380. ZHAO C H, ZHOU Y.Ship target detection and recognition based on improved Faster R-CNN algorithm[J].Journal of Shenyang University(Natural Science Edition), 2018, 30(5):366-371, 380.(in Chinese)
[12] REDMON J, DIVVALA S, GIRSHICK R, et al.You only look once:unified, real-time object detection[EB/OL].[2020-07-10].https://arxiv.org/abs/1506.02640.
[13] REDMON J, FARHADI A.YOLOv3:an incremental improvement[EB/OL].[2020-07-10].https://arxiv.org/abs/1804.02767.
[14] HE K M, ZHANG X Y, REN S Q, et al.Deep residual learning for image recognition[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2016:770-778.
[15] HUNG K Y, ZHANG Z K, JIANG J M.Real-time image super-resolution using recursive depthwise separable convolution network[J].IEEE Access, 2019, 7:99804-99816.
[16] 彭秀平, 仝其胜, 林洪彬, 等.一种面向散乱点云语义分割的深度残差-特征金字塔网络框架[J].自动化学报, 2019, 39(1):1-10. PENG X P, TONG Q S, LIN H B, et al.A deep residual-feature pyramid network framework for semantic segmentation of scattered point clouds[J].Acta Automatica Sinica, 2019, 39(1):1-10.(in Chinese)
[17] TANG J L, WANG D.Weed identification based on K-means feature learning combined with convolutional neural network[J].Computers and Electronics in Agriculture, 2017, 135(1):63-70.
[18] EVERINGHAM M, GOOL L, WILLIAMS C K I, et al.The pascal visual object classes challenge[J].International Journal of Computer Vision, 2010, 88(2):303-338.
[19] WANG Z, BOVIK A C, SHEIKH H R, et al.Image quality assessment:from error visibility to structural similarity[J].IEEE Transactions on Image Processing, 2004, 13(4):600-612.
[20] 周志华.机器学习[M].北京:清华大学出版社, 2016. ZHOU Z H.Machine learning[M].Beijing:Tsinghua University Press, 2016.(in Chinese)
[21] DOLLAR P, WOJEK C, SCHIELE B, et al.Pedestrian detection:a benchmark[C]//Proceedings of IEEE Conference on Computer Vision and Patten Recognition.Washington D.C., USA:IEEE Press, 2009:304-311.
[22] REN S, HE K, GIRSHICK R, et al.Faster R-CNN:towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6):1137-1149.

选择文件类型/文献管理软件名称

选择包含的内容