《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (7): 2248-2255.DOI: 10.11772/j.issn.1001-9081.2021050831

• 多媒体计算与计算机仿真 • 上一篇    

基于改进RetinaNet的船舶检测算法

凡文俊1, 赵曙光1(), 郭力争2   

  1. 1.东华大学 信息科学与技术学院,上海 201620
    2.河南城建学院 计算机与数据科学学院,河南 平顶山 467036
  • 收稿日期:2021-05-20 修回日期:2021-12-15 接受日期:2021-12-29 发布日期:2022-03-08 出版日期:2022-07-10
  • 通讯作者: 赵曙光
  • 作者简介:凡文俊(1996—),男,湖北天门人,硕士研究生,主要研究方向:人工智能、目标检测
    郭力争(1975—),男,河南开封人,副教授,博士,主要研究方向:云计算的资源管理与调度、机器学习。
  • 基金资助:
    中央高校基本科研业务费专项资金学科交叉重点计划项目(2232020A?12)

Ship detection algorithm based on improved RetinaNet

Wenjun FAN1, Shuguang ZHAO1(), Lizheng GUO2   

  1. 1.College of Information Science and Technology,Donghua University,Shanghai 201620,China
    2.School of Computer and Data Science,Henan University of Urban Construction,Pingdingshan Henan 467036,China
  • Received:2021-05-20 Revised:2021-12-15 Accepted:2021-12-29 Online:2022-03-08 Published:2022-07-10
  • Contact: Shuguang ZHAO
  • About author:FAN Wenjun, born in 1996, M. S. candidate. His research interests include artificial intelligence, target detection.
    GUO Lizheng, born in 1975, Ph. D., associate professor. His research interests include resource management and scheduling of cloud computing, machine learning.
  • Supported by:
    Interdisciplinary Key Program of Fundamental Research Funds for Central Universities(2232020A-12)

摘要:

目前基于深度学习算法的目标检测技术在合成孔径雷达(SAR)图像船舶检测中取得了显著的成果,然而仍存在着小目标船舶和近岸密集排列船舶检测效果差的问题。针对上述问题,提出了基于改进RetinaNet的船舶检测算法。在传统RetinaNet算法的基础上,首先,将特征提取网络残差块中的卷积改进为分组卷积,以增加网络宽度,从而提高网络的特征提取能力;其次,在特征提取网络的后两个阶段加入注意力机制,让网络更加专注于目标区域,从而提升目标检测能力;最后,将软非极大值抑制(Soft-NMS)加入到算法中,降低算法对于近岸密集排列船舶检测的漏检率。在高分辨率SAR图像数据集(HRSID)和SAR船舶检测数据集(SSDD)上的实验结果表明,所提改进算法对于小目标船舶和近岸船舶的检测效果得到了有效提升,与当前优秀的目标检测模型Faster R-CNN、YOLOv3和CenterNet等相比,在检测精度和速度上更加优越。

关键词: 合成孔径雷达图像, 船舶检测, RetinaNet, 注意力机制, 分组卷积

Abstract:

At present, the target detection technology based on deep learning algorithm has achieved the remarkable results in ship detection of Synthetic Aperture Radar (SAR) images. However, there is still the problem of poor detection effect of small target ships and densely arranged ships near shore. To solve the above problem, a new ship detection algorithm based on improved RetinaNet was proposed. On the basis of traditional RetinaNet algorithm, firstly, the convolution in the residual block of feature extraction network was improved to grouped convolution, thereby increasing the network width and improving the feature extraction ability of the network. Then, the attention mechanism was added in the last two stages of feature extraction network to make the network more focus on the target area and improve the target detection ability. Finally, the Soft Non-Maximum Suppression (Soft-NMS) was added to the algorithm to reduce the missed detection rate of the algorithm for the detection of densely arranged ships near shore. Experimental results on High-Resolution SAR Images Dataset (HRSID) and SAR Ship Detection Dataset (SSDD) show that, the proposed algorithm effectively improves the detection effect of small target ships and near-shore ships, is superior in detection precision and speed compared with the current excellent object detection models such as Faster Region-based Convolutional Neural Network (R-CNN), You Only Look Once version 3 (YOLOv3) and CenterNet.

Key words: Synthetic Aperture Radar (SAR) image, ship detection, RetinaNet, attention mechanism, grouped convolution

中图分类号: