Detection of oil spill based on CBF-CNN using HY-1C CZI multispectral images

Kai Du Yi Ma Zongchen Jiang Xiaoqing Lu Junfang Yang

Kai Du, Yi Ma, Zongchen Jiang, Xiaoqing Lu, Junfang Yang. Detection of oil spill based on CBF-CNN using HY-1C CZI multispectral images[J]. Acta Oceanologica Sinica, 2022, 41(7): 166-179. doi: 10.1007/s13131-021-1977-x
Citation: Kai Du, Yi Ma, Zongchen Jiang, Xiaoqing Lu, Junfang Yang. Detection of oil spill based on CBF-CNN using HY-1C CZI multispectral images[J]. Acta Oceanologica Sinica, 2022, 41(7): 166-179. doi: 10.1007/s13131-021-1977-x

doi: 10.1007/s13131-021-1977-x

Detection of oil spill based on CBF-CNN using HY-1C CZI multispectral images

Funds: The National Natural Science Foundation of China under contract No. 61890964; the Joint Funds of the National Natural Science Foundation of China under contract No. U1906217.
More Information
    Corresponding author: mayimail@fio.org.cn
    • 关键词:
    •  / 
    •  / 
    •  / 
    •  / 
    •  / 
    •  
  • Figure  1.  Study area 1. a. Geographical location of study area 1; b. the true color RGB image of study area 1; c. the angle between the viewing direction and the direction of mirror reflection (${\mathit{\theta }}_{\rm{m}}$) of study area 1; d. the spectra of oil emulsions, oil slicks, and seawater.

    Figure  2.  Study area 2 and study area 3. a. Geographical location of study area 1 and study area 2; b. the true color RGB image of area 2; c. the true color RGB image of area 3; d. the ${\theta }_{{\rm{m}}}$ of study area 2; e. the ${\theta }_{{\rm{m}}}$ of study area 3.

    Figure  3.  Rayleigh-corrected reflectance (Rrc) of study area 2 (a) and study area 3 (b).

    Figure  4.  Process of data augmentation.

    Figure  5.  Convolution neural networks structure.

    Figure  6.  Classification accuracy of oil spill under different sizes of spatial neighborhoods.

    Figure  7.  The time cost of different spatial neighborhood sizes.

    Figure  8.  Influence of data augmentation on the classification accuracy of oil emulsions and oil slicks. a. F1-Score of oil emulsions. b. F1-Score of oil slicks.

    Figure  9.  Influence of the values of $ \mathit{\alpha } $ and $ \mathit{\beta } $ on oil spill classification accuracy. a. The F1-Score of oil emulsions. b. The F1-Score of oil slicks.

    Figure  10.  Comparison of the results of the different loss functions. CBF: Class-Balanced F.

    Figure  11.  F1-Score of oil emulsions and oil slicks comparisons in different methods. CBF-CNN: Class-Balanced F convolution neural network; DNN: deep neural network; SVM: support vector machine; RF: Random Forests.

    Figure  12.  Classification results of different methods. a. Result of visual interpretation; b. result of the CBF-CNN model; c. result of the DNN model; d. result of the SVM; e. result of the RF model.

    Figure  13.  Spatial distribution of training data (a) and test data (b).

    Figure  14.  F1-Score of oil emulsions and oil slicks comparisons in different methods. CBF-CNN: Class-Balanced F convolution neural network; DNN: deep neural network; SVM: support vector machine; RF: Random Forests.

    Figure  15.  The results of different classification methods. a. Classification result of the CBF-CNN; b. classification result of the DNN; c. classification result of the SVM; d. classification result of the RF.

    Figure  16.  The logarithmic transformation of the image. a. Origin image; b. transformed image.

    Figure  17.  Training (a) and test data (b) of study area 3.

    Figure  18.  The results of different classification methods in study area 3. a. Classification result of the CBF-CNN; b. classification result of the DNN; c. classification result of the SVM; d. classification result of the RF.

    Figure  19.  Various oil spills could be identified from HY-1C CZI RGB images.

    Table  1.   Main technical specifications of the Haiyang-1C Coastal Zone Imager

    Band/nmCentral
    wavelength/nm
    Spatial
    resolution/m
    Signal-noise
    radio/dB
    420–50046050410
    520–60056050300
    610–69065050248
    760–89082550240
    下载: 导出CSV

    Table  2.   Parameters of different models

    ModelsParameters
    CBF-CNNEpoch: 100; Batch size: 450; Optimizer: Adam; Loss function: CBF; Learning rate: 0.001; Spatial neighborhood-scale: 11×11
    CNNEpoch: 100; Batch size: 450; Optimizer: Adam; Loss function: cross entropy; Learning rate: 0.001; Spatial neighborhood-scale: 11×11
    DNNHidden layer sizes: 50; Activation: ReLU; Optimizer: Adam; Learning rate: 0.2; Epoch:
    SVMKernel: RBF; C: 700; Gamma: 0.25; Degree: 3
    RFThe number of the trees: 90; The minimum number of samples on leaf: 50; The maximum number of elements: 5
    下载: 导出CSV

    Table  3.   Confusion matrices of different methods

    MethodsClassSeawaterOil emulsionsOil slicks
    CBF loss-CNNseawater699 12311 667
    oil emulsion1708 7041 576
    oil slick2 85784954 021
    DNNseawater700 30325463
    oil emulsion5277 9431 980
    oil slick23 06086933 798
    SVMseawater682 3609718 334
    oil emulsion2818 0542 115
    oil slick9 6891 12446 914
    RFseawater688 6201 89810 273
    oil emulsion1988 3871 865
    oil slick10 7261 66445 337
    下载: 导出CSV

    Table  4.   Statistics of training data and test data

    ClassOil emulsionsOil slicksBackgroundTotal
    Train28836912 48513 142
    Test9241 00031 31133 235
    Total1 2121 36943 79646 377
    下载: 导出CSV

    Table  5.   Parameters of different models

    ModelsParameters
    CBF-CNNEpoch: 100; Batch size: 450; Optimizer: Adam; Loss function: CBF; Learning rate: 0.001; Spatial neighborhood-scale: 11×11
    DNNHidden layer sizes: (100, 50); Activation: ReLU; Optimizer: Adam; Learning rate: 0.2; Epoch: 200
    SVMKernel: RBF; C: 500; Gamma: 0.5; Degree: 3
    RFThe number of the trees: 100; The minimum number of samples on leaf: 50; The maximum number of elements: 5
    下载: 导出CSV

    Table  6.   Statistics of train data and test data

    ClassOil slicksBackgroundTotal
    Training4 11110 58414 695
    Test7 79017 78225 572
    Total11 90128 36640 267
    下载: 导出CSV

    Table  7.   Parameters of different models

    ModelsParameters
    CBF-CNNEpoch: 100; Batch size: 450; Optimizer: Adam; Loss function: CBF; Learning rate: 0.001; Spatial neighborhood-scale: 11×11
    DNNHidden layer sizes: (100, 70); Activation: ReLU; Optimizer: Adam; Learning rate: 0.2; Epoch: 200
    SVMKernel: RBF; C: 800; Gamma: 0.35; Degree: 3
    RFThe number of the trees: 80; The minimum number of samples on leaf: 20; The maximum number of elements: 3
    下载: 导出CSV

    Table  8.   Detection performance of different methods for oil slicks

    MethodsPrecisionRecallF1-Score
    CBF-CNN0.940.970.96
    DNN0.690.840.76
    SVM0.660.860.75
    RF0.700.760.73
    下载: 导出CSV
  • [1] Abbriano R M, Carranza M M, Hogle S L, et al. 2011. Deepwater horizon oil spill: a review of the planktonic response. Oceanography, 24(3): 294–301. doi: 10.5670/oceanog.2011.80
    [2] Adamo M, De Carolis G, De Pasquale V, et al. 2009. Detection and tracking of oil slicks on sun-glittered visible and near infrared satellite imagery. International Journal of Remote Sensing, 30(24): 6403–6427. doi: 10.1080/01431160902865772
    [3] Breiman L. 2001. Random forests. Machine Learning, 45(1): 5–32. doi: 10.1023/A:1010933404324
    [4] Chen Weitao, Li Xianju, He Haixia, et al. 2018. A review of fine-scale land use and land cover classification in open-pit mining areas by remote sensing techniques. Remote Sensing, 10(1): 15
    [5] Corucci L, Nardelli F, Cococcioni M. 2010. Oil spill classification from multi-spectral satellite images: exploring different machine learning techniques. In: Proceedings of SPIE 7825, Remote Sensing of the Ocean, Sea Ice, and Large Water Regions 2010. Toulouse: SPIE, 782509
    [6] Cui Yin, Jia Menglin, Lin T Y, et al. 2019. Class-balanced loss based on effective number of samples. In: Proceedings of 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Long Beach, CA: IEEE
    [7] Esbaugh A J, Mager E M, Stieglitz J D. et al 2016. The effects of weathering and chemical dispersion on Deepwater Horizon crude oil toxicity to mahi-mahi (Coryphaena hippurus) early life stages. Science of The Total Environment, 543: 644–651. doi: 10.1016/j.scitotenv.2015.11.068
    [8] Feng Lian, Hou Xuejiao, Li Junsheng, et al. 2018. Exploring the potential of Rayleigh-corrected reflectance in coastal and inland water applications: a simple aerosol correction method and its merits. ISPRS Journal of Photogrammetry and Remote Sensing, 146: 52–64. doi: 10.1016/j.isprsjprs.2018.08.020
    [9] Hu Chuanmin. 2009. A novel ocean color index to detect floating algae in the global oceans. Remote Sensing of Environment, 113(10): 2118–2129. doi: 10.1016/j.rse.2009.05.012
    [10] Hu Chuanmin, Li Xiaofeng, Pichel W G, et al. 2009. Detection of natural oil slicks in the NW Gulf of Mexico using MODIS imagery. Geophysical Research Letters, 36(1): L01604
    [11] Hu Chuanmin, Lu Yingcheng, Sun Shaojie, et al. 2021. Optical remote sensing of oil spills in the ocean: what is really possible?. Journal of Remote Sensing, 2021: 9141902
    [12] Ioffe S, Szegedy C. 2015. Batch normalization: accelerating deep network training by reducing internal covariate shift. In: Proceedings of the 32nd International Conference on Machine Learning. Lille: ACM, 448–456
    [13] Jiang Zongchen, Ma Yi. 2020. Accurate extraction of offshore raft aquaculture areas based on a 3D-CNN model. International Journal of Remote Sensing, 41(14): 5457–5458. doi: 10.1080/01431161.2020.1737340
    [14] Jiang Zongchen, Ma Yi, Yang Junfang. 2020. Inversion of the thickness of crude oil film based on an OG-CNN Model. Journal of Marine Science and Engineering, 8(9): 653. doi: 10.3390/jmse8090653
    [15] Kingma D, Ba J. 2015. Adam: a method for stochastic optimization. In: Proceedings of the 3rd International Conference on Learning Representations. San Diego, CA: arXiv.org
    [16] Kolokoussis P, Karathanassi V. 2018. Oil spill detection and mapping using sentinel 2 imagery. Journal of Marine Science and Engineering, 6(1): 4. doi: 10.3390/jmse6010004
    [17] LeCun Y, Bengio Y. 1995. Convolutional networks for images, speech, and time series. In: Arbib M A, ed. The Handbook of Brain Theory and Neural Networks. Cambridge: MIT Press
    [18] Lin T Y, Goyal P, Girshick R, et al. 2017. Focal loss for dense object detection. In: Proceedings of 2017 IEEE International Conference on Computer Vision. Venice: IEEE, 2999–3007
    [19] Lu Yingcheng, Li Xiang, Tian Qingjiu, et al. 2013. Progress in marine oil spill optical remote sensing: detected targets, spectral response characteristics, and theories. Marine Geodesy, 36(3): 334–346. doi: 10.1080/01490419.2013.793633
    [20] Lu Yingcheng, Shi Jing, Hu Chuanmin, et al. 2020. Optical interpretation of oil emulsions in the ocean–Part II: Applications to multi-band coarse-resolution imagery. Remote Sensing of Environment, 242: 111778. doi: 10.1016/j.rse.2020.111778
    [21] Lu Yingcheng, Sun Shaojie, Zhang Minwei, et al. 2016. Refinement of the critical angle calculation for the contrast reversal of oil slicks under sunglint. Journal of Geophysical Research: Oceans, 121(1): 148–161. doi: 10.1002/2015JC011001
    [22] Lu Jinshu, Xu Zhenfeng, Xu Song, et al. 2015. Experimental and numerical investigations on reliability of air barrier on oil containment in flowing water. Marine Pollution Bulletin, 95(1): 200–206. doi: 10.1016/j.marpolbul.2015.04.020
    [23] Michel J, Owens E H, Zengel S, et al. 2013. Extent and degree of shoreline oiling: Deepwater horizon oil spill, Gulf of Mexico, USA. PLoS ONE, 8(6): e65087. doi: 10.1371/journal.pone.0065087
    [24] Niclòs R, Doña C, Valor E, et al. 2013. Thermal-infrared spectral and angular characterization of crude oil and seawater emissivities for oil slick identification. IEEE Transactions on Geoscience and Remote Sensing, 52(9): 5387–5395
    [25] Serra-Sogas N, O’Hara P D, Canessa R, et al. 2008. Visualization of spatial patterns and temporal trends for aerial surveillance of illegal oil discharges in western Canadian marine waters. Marine Pollution Bulletin, 56(5): 825–833. doi: 10.1016/j.marpolbul.2008.02.005
    [26] Shen Yafeng, Liu Jianqiang, Ding Jing, et al. 2020. HY-1C COCTS and CZI observation of marine oil spills in the South China Sea. Journal of Remote Sensing, 24(8): 933–944
    [27] Sun Shaojie, Lu Yingcheng, Liu Yongxue, et al. 2018. Tracking an oil tanker collision and spilled oils in the East China Sea using multisensor day and night satellite imagery. Geophysical Research Letters, 45(7): 3212–3220. doi: 10.1002/2018GL077433
    [28] Tong Cheng, Mu Bing, Liu Rongjie, et al. 2019. Atmospheric correction algorithm for HY-1C CZI over turbid waters. Journal of Coastal Research, 90(SI): 156–163
    [29] Wen Yansha, Wang Mengqiu, Lu Yingcheng, et al. 2018. An alternative approach to determine critical angle of contrast reversal and surface roughness of oil slicks under sunglint. International Journal of Digital Earth, 11(9): 972–979. doi: 10.1080/17538947.2018.1470687
    [30] Yang Junfang, Wan Jianhua, Ma Yi, et al. 2019. Oil spill hyperspectral remote sensing detection based on DCNN with multi-scale features. Journal of Coastal Research, 90(SI): 332–339
    [31] Yekeen S T, Balogun A L, Yusof K B W. 2020. A novel deep learning instance segmentation model for automated marine oil spill detection. ISPRS Journal of Photogrammetry and Remote Sensing, 167: 190–200. doi: 10.1016/j.isprsjprs.2020.07.011
    [32] Yin Liping, Zhang Min, Zhang Yuanling, et al. 2018. The long-term prediction of the oil-contaminated water from the Sanchi collision in the East China Sea. Acta Oceanologica Sinica, 37(3): 69–72. doi: 10.1007/s13131-018-1193-5
    [33] Zhu Xueyuan, Li Ying, Zhang Qiang, et al. 2019. Oil film classification using deep learning-based hyperspectral remote sensing technology. ISPRS International Journal of Geo-Information, 8(4): 181. doi: 10.3390/ijgi8040181
  • 加载中
图(19) / 表(8)
计量
  • 文章访问数:  470
  • HTML全文浏览量:  161
  • PDF下载量:  28
  • 被引次数: 0
出版历程
  • 收稿日期:  2021-03-04
  • 录用日期:  2021-11-12
  • 网络出版日期:  2022-01-25
  • 刊出日期:  2022-07-08

目录

    /

    返回文章
    返回