سابِق|تال

JAM-R-CNN deep learning network model for remote sensing recognition of terraced fields

  • role： First author第一作者
  • Affiliation：

    College of Urban and Environmental Sciences, Central China Normal University, Wuhan 430079, China

    Hubei Province Key Laboratory for Geographical Process Analysis and Simulation, Wuhan 430079, China

  • Email：xjy959@mails.ccnu.edu.cn
  • Introduction：谢君洋，研究方向为农业遥感。E-mail： xjy959@mails.ccnu.edu.cn
XIE Junyang12,  
  • Affiliation：

    College of Urban and Environmental Sciences, Central China Normal University, Wuhan 430079, China

    Hubei Province Key Laboratory for Geographical Process Analysis and Simulation, Wuhan 430079, China

LIN Anqi12,  
  • role： Corresponding author通信作者
  • Affiliation：

    College of Urban and Environmental Sciences, Central China Normal University, Wuhan 430079, China

    Hubei Province Key Laboratory for Geographical Process Analysis and Simulation, Wuhan 430079, China

  • Email：haowu@ccnu.edu.cn
  • Introduction：吴浩，研究方向为地理信息科学与遥感。E-mail： haowu@ccnu.edu.cn
WU Hao12*,  
  • Affiliation：

    College of Urban and Environmental Sciences, Central China Normal University, Wuhan 430079, China

    Hubei Province Key Laboratory for Geographical Process Analysis and Simulation, Wuhan 430079, China

WU Ziwei12,  
  • Affiliation：

    Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, China

WU Wenbin3,  
  • Affiliation：

    Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, China

YU Qiangyi3

ملخص

Efficiently and accurately determining the spatial distribution of terraced fields provides important data support for soil and water conservation and improves the regulatory level of agriculture in mountainous areas. When deep learning methods are used for terrace recognition, narrow and elongated terraces are prone to be missed because of convolution operations, and in complex backgrounds with unclear terrain boundaries in mountainous areas, large areas of adhesive recognition results are easily generated, leading to low accuracy in the final terrace recognition. Prior to the achievement of accurate recognition of terrace information, the urgent technical problems to be solved are how to effectively maintain the high semantic information of high-resolution remote sensing images in the convolution operation process on the basis of the characteristics of terraces and how to reduce the omission of narrow and long terraces and the adhesion of recognition results. To address these problems, this study proposes the JAM-R-CNN deep learning network terrace recognition method that adopts remote sensing images with very high resolution. This network is based on the Mask Region-based Convolutional Neural Network (Mask R-CNN) model. It integrates the jumping network to maintain the high semantic information of high-resolution remote sensing images, employs the convolutional block attention module to enhance the feature expression ability of terraces, and modifies the anchor size to adapt to the narrow and long characteristics of terraces and improve terrace recognition accuracy. A part of the salt well terraces in Nanchuan District, Chongqing, China, is selected as the study area to test the proposed method, and four models in domestic GF-2 satellite image data are used for experiments. Results show that the terrace parcel map derived from the JAM-R-CNN model has a precision of 90.81%, recall of 84.28%, F1 score of 88.98%, and Intersection over Union (IoU) value of 83.15%. Compared with Mask R-CNN, JAM-R-CNN’s precision, recall, F1 score, and IoU value are increased by 1.96%, 5.26%, 3.29%, and 5.19%, respectively, indicating that the JAM-R-CNN model can better identify the terraces than Mask R-CNN can. Most of the terraces identified by Unet and DeepLab v3+ are connected together, and the terraced fields with a small size are not distinguished. The JAM-R-CNN model identifies fewer missing areas on the periphery of terraces compared with Mask R-CNN model, and the number of missing narrow and long terraces is considerably reduced. This result is the effect of three improved parts and further proves that the proposed JAM-R-CNN model exerts a remarkable improvement effect and demonstrates superior performance in remote sensing recognition of terraces. The proposed JAM-R-CNN deep learning network model effectively reduces the adhesion phenomenon of terrace recognition results and considerably improves the extraction rate for narrow and long terraces, thus achieving a substantial improvement in the overall accuracy of terrace remote sensing recognition. The model has good application value.

مفهوم

remote sensing;terrace recognition;high resolution remote sensing image;deep learning;jump network;JAM-R-CNN

References

  1. 1.
    Cai Z W, He Z, Wang W J, Yang J Y, Wei H D, Wang C and Xu B D. 2022. Mapping cropland at metric resolution using the spatiotemporal information from multi-source GF satellite data. National Remote Sensing Bulletin, 26(7): 1368-1382
  2. 2.
    Chen L L, Shi Z, Liao K T, Song Y J and Zhang H M. 2022. Cultivated land extraction from high resolution remote sensing image based on convolutional neural network. Transactions of the Chinese Society for Agricultural Machinery, 53(9): 168-177
  3. 3.
    Dai W, Na J M, Huang N, Hu G H, Yang X, Tang G A, Xiong L Y and Li F Y. 2020. Integrated edge detection and terrain analysis for agricultural terrace delineation from remote sensing images. International Journal of Geographical Information Science, 34(3): 484-503.
  4. 4.
    Dang T M, Mu X M, Sun W Y, Gao P and Zhao G J. 2017. Review of quickly discriminating approaches of terrace information based on high resolution remote sensing images. Yellow River, 39(3): 85-89, 94
  5. 5.
    Deng P F, Xu K J and Huang H. 2021. CNN-GCN-based dual-stream network for scene classification of remote sensing images. National Remote Sensing Bulletin, 25(11): 2270-2282
  6. 6.
    Diaz-Varela R A, Zarco-Tejada P J, Angileri V and Loudjani P. 2014. Automatic identification of agricultural terraces through object-oriented analysis of very high resolution DSMs and multispectral imagery obtained from an unmanned aerial vehicle. Journal of Environmental Management, 134: 117-126.
  7. 7.
    Dong S J, Xin L J, Li S F, Xie H L, Zhao Y L, Wang X, Li X B, Song H F and Lu Y H. 2023. The extent and spatial distribution of terrace abandonment in China. Acta Geographica Sinica, 78(1): 3-15
  8. 8.
    Eckert S, Ghebremicael S T, Hurni H and Kohler T. 2017. Identification and classification of structural soil conservation measures based on very high resolution stereo satellite data. Journal of Environmental Management, 193: 592-606.
  9. 9.
    He K M, Gkioxari G, Dollár P and Girshick R. 2017. Mask R-CNN//International Conference on Computer Vision. Venice: IEEE: 2980-2988.
  10. 10.
    Jong M, Guan K Y, Wang S B, Huang Y Z and Peng B. 2022. Improving field boundary delineation in ResUNets via adversarial deep learning. International Journal of Applied Earth Observation and Geoinformation, 112: 102877.
  11. 11.
    Liu W, Wu Z F, Luo J C, Sun Y W, Wu T J, Zhou N, Hu X D, Wang L Y and Zhou Z F. 2021. A divided and stratified extraction method of high-resolution remote sensing information for cropland in hilly and mountainous areas based on deep learning. Acta Geodaetica et Cartographica Sinica, 50(1): 105-116
  12. 12.
    Liu X Y, Yang S T, Wang F G, He X Z, Ma H B and Luo Y. 2014. Analysis on sediment yield reduced by current terrace and shrubs-herbs-arbor vegetation in the Loess Plateau. Journal of Hydraulic Engineering, 45(11): 1293-1300
  13. 13.
    Long J, Li M M, Wang X Q and Stein A. 2022. Delineation of agricultural fields using multi-task BsiNet from high-resolution satellite images. International Journal of Applied Earth Observation and Geoinformation, 112: 102871.
  14. 14.
    Luo L H, Li F Y, Dai Z Y, Yang X, Liu W and Fang X. 2020. Terrace extraction based on remote sensing images and digital elevation model in the loess plateau, China. Earth Science Informatics, 13(2): 433-446.
  15. 15.
    Mei W Y, Wang H Y, Fouhey D, Zhou W Q, Hinks I, Gray J M, Van Berkel D and Jain M. 2022. Using deep learning and very-high-resolution imagery to map smallholder field boundaries. Remote Sensing, 14(13): 3046.
  16. 16.
    Sun W Y, Zhang Y Q, Mu X M, Li J Y, Gao P, Zhao G J, Dang T M and Chiew F. 2019. Identifying terraces in the hilly and gully regions of the Loess Plateau in China. Land Degradation & Development, 30(17): 2126-2138.
  17. 17.
    Waldner F and Diakogiannis F I. 2020. Deep learning on edge: Extracting field boundaries from satellite images with a convolutional neural network. Remote Sensing of Environment, 245: 111741.
  18. 18.
    Wang J D, Sun K, Cheng T H, Jiang B R, Deng C R, Zhao Y, Liu D, Mu Y D, Tan M K, Wang X G, Liu W Y and Xiao B. 2021. Deep high-resolution representation learning for visual recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(10): 3349-3364.
  19. 19.
    Woo S, Park J, Lee J Y and Kweon I S. 2018. CBAM: convolutional block attention module//15th European Conference on Computer Vision. Munich: Springer: 3-19.
  20. 20.
    Xu Z Y, Shen Z F, Li Y, Zhao L F, Ke Y M, Li L L and Wen Q. 2022. Classification of high-resolution remote sensing images based on enhanced DeepLab algorithm and adaptive loss function. National Remote Sensing Bulletin, 26(2): 406-415
  21. 21.
    Xue M D, Zhang H M, Yang J T and Li X S. 2019. Extraction of terrace information based on UAV image and topographic index. Application Research of Computers, 36(8): 2527-2533, 2538
  22. 22.
    Yu M G, Rui X P, Xie W Y, Xu X J and Wei W. 2022. Research on automatic identification method of terraces on the loess plateau based on deep transfer learning. Remote Sensing, 14(10): 2446.
  23. 23.
    Yuan S, Tang H and Guo Y. 2022. Segmentation method for plant leaves using an improved Mask R-CNN model. Transactions of the Chinese Society of Agricultural Engineering, 38(1): 212-220
  24. 24.
    Zhao F, Xiong L Y, Wang C, Wang H R, Wei H and Tang G A. 2021. Terraces mapping by using deep learning approach from remote sensing images and digital elevation models. Transactions in GIS, 25(5): 2438-2454.
  25. 25.
    Zhao X, Wang X J, Zhao Y, Luo Z D, Xu Y L, Guo H and Zhang Y. 2016. Feasibility analysis on methodology of terraced extraction by using Fourier transformation based on domestic hi-resolution remote sensing images of GF-1 satellite. Soil and Water Conservation in China, (1): 63-65
  26. 26.
    Zhou N, Yang P, Wei C S, Shen Z F, Yu J J, Ma X Y and Luo J C. 2021. Accurate extraction method for cropland in mountainous areas based on field parcel. Transactions of the Chinese Society of Agricultural Engineering, 37(19): 260-266
  27. 27.
    Zhou Y, Li M M, Wang X Q, Wu S Y and Jin S L. 2022. Extraction of farming terraces using object-based convolutional neural networks from very high resolution satellite images. Remote Sensing Information, 37(2): 138-144

قراءة النص الكامل

The above content is generated by Large Model Translation. The translated content is for reference only. We do not assume any commercial or legal responsibilty for any consequences arising from the use of our website