面向小型船舶遥感智能识别的改进型YOLOv5s模型
An Improved YOLOv5s Model for Intelligent Recognition of Small Ships with Remote Sensing
DOI: 10.12677/gst.2024.123031, PDF,   
作者: 陈春晓：四川测绘地理信息局测绘技术服务中心，四川 成都
关键词: 光学遥感；深度学习；小型船舶；智能识别；改进模型；Optical Remote Sensing； Deep Learning； Small Ships； Intelligent Recognition； Improved Model
摘要: 船舶探测在领土安全和海洋环境保护方面具有重要意义。然而，在复杂环境中准确地检测识别小型船舶仍是存在一定技术难度与挑战。考虑到小型船舶在遥感图像上的特征信息较少，本研究提出了一种基于YOLOv5s的小型船舶检测模型。该模型首先在顶端添加了浅层特征图的检测层，并在中部采用了跳过连接，用以提高小型船只的检测精度；随后提出了一种新颖有效的混合空间金字塔池化(Hybrid Spatial Pyramid Pooling, HSPP)技术，以融合特征图的局部和全局信息；此外，在骨干中采用坐标注意(Coordinate Attention, CA)机制来增强小型船舶的表征，并使用高效IOU (EIOU)作为边界框回归的损失函数，以提高所提模型的定位精度；最终，使用K-means++算法来获得更合理的小型船舶检测锚点。实验选取包括10,133幅图像的多尺度船舶数据集OSSD和公开船舶数据集LEVIR-Ship，结合同类既有模型开展了比较分析，验证模型和算法的有效性。
Abstract: Ship detection is of great importance in territorial security and marine environmental protection. However, the accurate detection of small ships is a challenging task in complex environments mainly due to small ships having few features on remote sensing images. In this paper, we propose a small ship detection model based on YOLOv5s. Firstly, a detection layer is added with a shallow feature map in the head and skip connections are employed in the neck to improve the detection accuracy of small ships. Then, a novel and effective hybrid spatial pyramid pooling (Hybrid Spatial Pyramid Pooling, HSPP) is proposed to fuse the local and global information of feature maps. In addition, a coordinate attention (Coordinate Attention, CA) mechanism is employed in the backbone to augment the representations of small ships, and efficient IOU (EIOU) is used as the loss function for bounding box regression to enhance the localization accuracy of the proposed model. Finally, K-means++ algorithm is used to obtain more reasonable anchors for small ship detection. We introduce the multiscale ship dataset OSSD that contains 10,133 images, and LEVIR-Ship opening datasets. Experiments on LEVIR-Ship and OSSD validated the effectiveness of our proposed model.
文章引用：陈春晓. 面向小型船舶遥感智能识别的改进型YOLOv5s模型[J]. 测绘科学技术, 2024, 12(3): 246-256. https://doi.org/10.12677/gst.2024.123031

参考文献

[1] Proia, N. and Page, V. (2010) Characterization of a Bayesian Ship Detection Method in Optical Satellite Images. IEEE Geoscience and Remote Sensing Letters, 7, 226-230. [Google Scholar] [CrossRef] 
[2] Liu, Y., Zhao, J. and Qin, Y. (2021) A Novel Technique for Ship Wake Detection from Optical Images. Remote Sensing of Environment, 258, Article ID: 112375. [Google Scholar] [CrossRef] 
[3] Chen, J., Chen, K., Chen, H., Li, W., Zou, Z. and Shi, Z. (2022) Contrastive Learning for Fine-Grained Ship Classification in Remote Sensing Images. IEEE Transactions on Geoscience and Remote Sensing, 60, 1-16. [Google Scholar] [CrossRef] 
[4] Han, W., Kuerban, A., Yang, Y., Huang, Z., Liu, B. and Gao, J. (2022) Multi-vision Network for Accurate and Real-Time Small Object Detection in Optical Remote Sensing Images. IEEE Geoscience and Remote Sensing Letters, 19, 1-5. [Google Scholar] [CrossRef] 
[5] Zhang, Z., Guo, W., Zhu, S. and Yu, W. (2018) Toward Arbitrary-Oriented Ship Detection with Rotated Region Proposal and Discrimination Networks. IEEE Geoscience and Remote Sensing Letters, 15, 1745-1749. [Google Scholar] [CrossRef] 
[6] Xu, Q., Li, B., He, Z. and Ma, C. (2011) Multiscale Contour Extraction Using a Level Set Method in Optical Satellite Images. IEEE Geoscience and Remote Sensing Letters, 8, 854-858. [Google Scholar] [CrossRef] 
[7] Yang, G., Li, B., Ji, S.F., Gao, F. and Xu, Q.Z. (2014) Ship Detection from Optical Satellite Images Based on Sea Surface Analysis. IEEE Geoscience and Remote Sensing Letters, 11, 641-645. [Google Scholar] [CrossRef] 
[8] Li, Y., Zhang, S. and Wang, W. (2022) A Lightweight Faster R-CNN for Ship Detection in SAR Images. IEEE Geoscience and Remote Sensing Letters, 19, 1-5. [Google Scholar] [CrossRef] 
[9] Shorten, C. and Khoshgoftaar, T.M. (2019) A Survey on Image Data Augmentation for Deep Learning. Journal of Big Data, 6, Article No. 60. [Google Scholar] [CrossRef] 
[10] Chen, J., Chen, K., Chen, H., Zou, Z. and Shi, Z. (2022) A Degraded Reconstruction Enhancement-Based Method for Tiny Ship Detection in Remote Sensing Images with a New Large-Scale Dataset. IEEE Transactions on Geoscience and Remote Sensing, 60, 1-14. [Google Scholar] [CrossRef] 
[11] Lin, T., Dollar, P., Girshick, R., He, K., Hariharan, B. and Belongie, S. (2017) Feature Pyramid Networks for Object Detection. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, 21-26 July 2017, 936-944. [Google Scholar] [CrossRef] 
[12] Liu, S., Qi, L., Qin, H., Shi, J. and Jia, J. (2018) Path Aggregation Network for Instance Segmentation. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, 18-23 June 2018, 8759-8768. [Google Scholar] [CrossRef] 
[13] Tan, M., Pang, R. and Le, Q.V. (2020) Efficientdet: Scalable and Efficient Object Detection. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, 13-19 June 2020, 10778-10787. [Google Scholar] [CrossRef] 
[14] He, K.M., Zhang, X.Y., Ren, S.Q., et al. (2015) Spatial Pyramid Pooling in Deep Convolutional Net-Works for Visual Recognition. IEEE Transactions on Pattern Analysis & Machine Intelligence, 37, 1904-1916.
[15] Hou, Q., Zhou, D. and Feng, J. (2021) Coordinate Attention for Efficient Mobile Network Design. 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Nashville, 20-25 June 2021, 13708-13717. [Google Scholar] [CrossRef] 
[16] Zhang, Y., Ren, W., Zhang, Z., Jia, Z., Wang, L. and Tan, T. (2022) Focal and Efficient IOU Loss for Accurate Bounding Box Regression. Neurocomputing, 506, 146-157. [Google Scholar] [CrossRef] 
[17] Arthur, D. and Vassilvitskii, S. (2007) K-Means++: The Advantages of Careful Seeding. Proceedings of the Eighteenth Annual ACM-SIAM Symposium on Discrete Algorithms, SODA 2007, New Orleans, 7-9 January 2007.
https://dl.acm.org/doi/10.5555/1283383.1283494 .
[18] Shi, Q., Liu, M., Marinoni, A. and Liu, X. (2023) Ugs-1m: Fine-Grained Urban Green Space Mapping of 31 Major Cities in China Based on the Deep Learning Framework. Earth System Science Data, 15, 555-577. [Google Scholar] [CrossRef] 
[19] He, D., Shi, Q., Liu, X., Zhong, Y., Xia, G. and Zhang, L. (2022) Generating Annual High Resolution Land Cover Products for 28 Metropolises in China Based on a Deep Super-Resolution Mapping Network Using Landsat Imagery. GIScience & Remote Sensing, 59, 2036-2067. [Google Scholar] [CrossRef] 

为你推荐