Marine cage aquaculture information extraction based on SLA-UNet
-
摘要: 网箱养殖是海水养殖中最重要的类型之一，各类网箱在遥感影像中形状不一，且背景复杂，以往的网箱提取方法，未能完全模拟人类的视觉行为，以及高效利用光谱信息。针对上述问题，提出深度多循环注意力光谱的U-Net网络模型（Spectral Loopy Attention U-Net, SLA-UNet）进行网箱养殖信息提取，使用基于最优尺度寻优（Estimation of Scale Parameter, ESP）的随机森林（Random Forest, RF）算法，去除波段运算后的冗余光谱信息，并添加类似人眼的注意力行为机制，深化影响网箱信息提取的重要特征通道，同时进行边缘补齐补充损失信息，实现了网箱养殖信息的高精度提取。选取广东省湛江市和海南省临高县作为研究区域，与Canny算子、Otsu算法、PCA_Kmeans算法、基于ESP的RF算法、U-Net模型提取结果进行对比，所提SLA-UNet模型近岸网箱的提取精度为98.3%，深海网箱提取精度平均值为98.9%，验证了SLA-UNet模型在网箱养殖识别中的有效性。Abstract: Cage aquaculture is one of the most important types of marine aquaculture. Different types of cages have varying shapes in remote sensing images, and the background is complex. Previous methods for cage extraction have not been able to fully simulate human visual behavior and efficiently utilize spectral information. To address these issues, we propose a Spectral Loopy Attention U-Net (SLA-UNet) network model for cage aquaculture information extraction. The model utilizes the Random Forest (RF) algorithm based on the Estimation of Scale Parameter (ESP) to remove redundant spectral information after band operations. It also incorporates a human-like attention mechanism to enhance the important feature channels that affect cage information extraction. Additionally, edge completion is performed to supplement the loss information, achieving high-precision extraction of cage aquaculture information. We selected Zhanjiang City, Guangdong Province and Lingao County, as the study areas. Comparisons were made with the extraction results of the Canny algorithm, Otsu algorithm, PCA_Kmeans algorithm, RF algorithm based on ESP, and the U-Net model. The extraction accuracy of the SLA-UNet model for nearshore cages is 98.3%, and the average extraction accuracy for deep-sea cages is 98.9%, validating the effectiveness of the SLA-UNet model in cage aquaculture recognition.
-
图 1 湛江市坡头区和霞山区海水网箱养殖区
Fig. 1 Marine cage aquaculture area at Potou and Xiashan in Zhanjiang
图 3 海水网箱养殖
a. 传统近岸网箱；b. 新型深海网箱养殖
Fig. 3 Marine cage aquaculture
a. Traditional offshore cages; b. new deep-sea cages
表 1 近岸网箱精度验证结果
Tab. 1 Accuracy verification results of nearshore cage
测试区 Canny算子 Otsu算法 PCA_Kmeans算法 基于ESP的RF算法 U-Net SLA-UNet OA/% 84.7 87.3 88.2 90.5 95.6 98.3 R/% 86.3 89.2 90.3 91.5 97.8 98.6 MIOU/% 55.7 62.5 65.7 67.2 80.2 83.5 
下载: 导出CSV
表 2 湛江深海网箱精度验证结果
Tab. 2 Accuracy verification results of deep sea cage in Zhanjiang
测试区 Canny算子 Otsu算法 PCA_Kmeans算法 基于ESP的RF算法 U-Net SLA-UNet OA/% 85.2 86.7 88.9 91.3 97.3 99.2 R/% 85.4 88.1 88.5 89.3 97.5 98.3 MIOU/% 60.5 63.3 66.8 72.1 86.2 91.6
下载: 导出CSV
表 3 临高县深海网箱精度验证结果
Tab. 3 Accuracy verification results of deep sea cage at Lingao
测试区 Canny算子 Otsu算法 PCA_Kmeans算法 基于ESP的RF算法 U-Net SLA-UNet OA/% 75.1 79.3 80.2 85.6 95.1 98.6 R/% 78.4 80.2 82.5 86.2 95.6 98.1 MIOU/% 59.1 59.9 61.7 69.8 76.4 82.3
下载: 导出CSV
-
[1] Wang Ming, Mao Dehua, Xiao Xiangming, et al. Interannual changes of coastal aquaculture ponds in China at 10-m spatial resolution during 2016–2021[J]. Remote Sensing of Environment, 2023, 284: 113347. doi: 10.1016/j.rse.2022.113347 [2] FAO. The State of World Fisheries and Aquaculture 2020[EB/OL]. [2023-03-05]. http://doi.org/10.4060/ca9229en [3] Mao Dehua, Yang Hong, Wang Zongming, et al. Reverse the hidden loss of China’s wetlands[J]. Science, 2022, 376(6597): 1061. [4] 邓莎萨, 张帆, 尹嫱等. 面向目视解译的全极化SAR船只精细化特征表征方法[J/OL]. [2024-02-21]. 雷达学报, 2023. https://radars.ac.cn/cn/article/doi/ 10.12000/JR23078. doi: 10.12000/JR23078Deng Shasa, Zhang Fan, Yin Qiang, et al. Refined ship feature characterization method of full-polarimetric synthetic aperture radar for visual interpretation[J/OL]. Journal of Radars, 2023. https://radars.ac.cn/cn/article/doi/ 10.12000/JR23078. doi: 10.12000/JR23078 [5] Schepaschenko D, See L, Lesiv M, et al. Recent advances in forest observation with visual interpretation of very high-resolution imagery[J]. Surveys in Geophysics, 2019, 40(4): 839−862. doi: 10.1007/s10712-019-09533-z [6] 汪静平, 吴小丹, 马杜娟, 等. 基于机器学习的遥感反演: 不确定性因素分析[J]. 遥感学报, 2023, 27(3): 790−801.Wang Jingping, Wu Xiaodan, Ma Dujuan, et al. Remote sensing retrieval based on machine learning algorithm: uncertainty analysis[J]. Journal of Remote Sensing, 2023, 27(3): 790−801. [7] Virnodkar S S, Pachghare V K, Patil V C, et al. Remote sensing and machine learning for crop water stress determination in various crops: A critical review[J]. Precision Agriculture, 2020, 21(5): 1121−1155. doi: 10.1007/s11119-020-09711-9 [8] Maxwell A E, Warner T A, Guillén L A. Accuracy assessment in convolutional neural network-based deep learning remote sensing studies—Part 2: Recommendations and best practices[J]. Remote Sensing, 2021, 13(13): 2591. doi: 10.3390/rs13132591 [9] Kechagias-Stamatis O, Aouf N. Fusing deep learning and sparse coding for SAR ATR[J]. IEEE Transactions on Aerospace and Electronic Systems, 2019, 55(2): 785−797. doi: 10.1109/TAES.2018.2864809 [10] Seto K C, Fragkias M. Mangrove conversion and aquaculture development in Vietnam: A remote sensing-based approach for evaluating the Ramsar Convention on Wetlands[J]. Global Environmental Change, 2007, 17(3/4): 486−500. [11] Hinton G E, Salakhutdinov R R. Reducing the dimensionality of data with neural networks[J]. Science, 2006, 313(5786): 504−507. doi: 10.1126/science.1127647 [12] Long J, Shelhamer E, Darrell T. Fully convolutional Networks for semantic segmentation[C]//Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston: IEEE, 2015: 3431−3440. [13] Li Jiayi, Zhang Hongyan, Zhang Liangpei. Supervised segmentation of very high resolution images by the use of extended morphological attribute profiles and a sparse transform[J]. IEEE Geoscience and Remote Sensing Letters, 2014, 11(8): 1409−1413. doi: 10.1109/LGRS.2013.2294241 [14] Ronneberger O, Fischer P, Brox T. U-Net: Convolutional Networks for biomedical image segmentation[C]//Proceedings of the 18th International Conference on Medical Image Computing and Computer-Assisted Intervention. Munich: Springer, 2015: 234−241. [15] Badrinarayanan V, Kendall A, Cipolla R. SegNet: A deep convolutional encoder-decoder architecture for image segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(12): 2481−2495. doi: 10.1109/TPAMI.2016.2644615 [16] Zheng Shuai, Jayasumana S, Romera-Paredes B, et al. Conditional random fields as recurrent neural networks[C]//Proceedings of the 2015 IEEE International Conference on Computer Vision. Santiago: IEEE, 2015: 1529−1537. [17] Chen Yang, Fan Rongshuang, Yang Xiucheng, et al. Extraction of urban water bodies from high-resolution remote-sensing imagery using deep learning[J]. Water, 2018, 10(5): 585. doi: 10.3390/w10050585 [18] 刘岳明, 杨晓梅, 王志华, 等. 基于深度学习RCF模型的三都澳筏式养殖区提取研究[J]. 海洋学报, 2019, 41(4): 119−130.Liu Yueming, Yang Xiaomei, Wang Zhihua, et al. Extracting raft aquaculture areas in Sanduao from high-resolution remote sensing images using RCF[J]. Haiyang Xuebao, 2019, 41(4): 119−130. [19] Liu Chenxi, Jiang Tao, Zhang Zhen, et al. Extraction method of offshore mariculture area under weak signal based on multisource feature fusion[J]. Journal of Marine Science and Engineering, 2020, 8(2): 99. doi: 10.3390/jmse8020099 [20] 柯丽娜, 翟宇宁, 范剑超. 深度边缘光谱U-Net海水网箱养殖信息提取[J]. 海洋学报, 2022, 44(2): 132−142.Ke Lina, Zhai Yuning, Fan Jianchao. Marine cage aquaculture information extraction based on deep edge spectral U-Net[J]. Haiyang Xuebao, 2022, 44(2): 132−142. -
图(11) / 表(3)
计量
- 文章访问数: 55
- HTML全文浏览量: 32
- PDF下载量: 7
- 被引次数: 0
