Anterior|Siguiente

Topographic correction effect on classification accuracy for deep neural network classifier—A case study of the U-Net model

  • role： First author第一作者
  • Affiliation：

    Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China

    University of Chinese Academy of Sciences, School of Electronic, Electrical and Communication Engineering, Beijing 100049, China

  • Email：jiali_199311@163.cm
  • Introduction：贾莉，研究方向为遥感图像处理、地表覆盖分类。E-mail：jiali_199311@163.cm
JIA Li12,  
  • Affiliation：

    Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China

ZHENG Ke1,  
  • Affiliation：

    Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China

TANG Ping1,  
  • role： Corresponding author通信作者
  • Affiliation：

    Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China

  • Email：huolz@radi.ac.cn
  • Introduction：霍连志，研究方向为遥感图像处理、遥感图像分类技术。E-mail： huolz@radi.ac.cn
HUO Lianzhi1*

resumen

Remote sensing image classification technology has been widely used in land resource monitoring, forest resource investigation, and other related fields. For the traditional classification methods, terrain effect is an unavoidable factor that restricts the improvement of classification accuracy in the classification of land cover types, and the influence can be weakened by an appropriate correction model, and topographic correction has been proven to play a positive role in improving classification accuracy. Compared with traditional classifiers, deep neural network classifiers based on deep learning theory have the advantages of deep feature learning and expression, which have emerged in the field of image classification and have been gradually applied to land cover classification with good results.Our study has some shortcomings. First, some errors remain in GlobeLand30 and national forest classification products compared with the actual surface cover types. Therefore, more detailed and accurate classification results need to be found as sample labels in the future. Second, the surface of sunny-shady slopes with different types of coverage is not considered in this study. Finally, U-Net should not be the sole focus. In subsequent research, we will attempt to select a variety of deep neural network classifiers to explore the classification accuracy changes before and after topographic correction of neural network models at a more precise scale to obtain further conclusions.This paper further explores the influence of topographic correction on the classification accuracy of land cover classification by deep neural network classifiers.Using Landsat 8 OLI image data and GDEM_V2 terrain data with 30 m resolution as data sources, and based on the classification results of GlobeLand30 and national forest types, this paper implemented the classification extraction of land cover types by the U-Net semantic segmentation network and made a comparative analysis of classification accuracy before and after topographic correction with different sample acquisition methods and different levels of classification systems.Classification results show the following. (1) With two training sample acquisition methods, namely, grid clipping and aspect auxiliary clipping, the classification accuracy after topographic correction is unchanged or slightly reduced compared with that before correction, and the reduction range is 0.9%—1.39%. (2) For the classification of more precise forest types, the classification accuracy after topographic correction decreased by 1.66% compared with that before correction.In this paper, we conduct a preliminary study and find that in the classification of land cover types by the U-Net model with different sample acquisition methods, namely, regular grid clipping and aspect auxiliary clipping, and under different classification systems, the classification accuracy of the deep neural network classifier-U-Net is not improved by the topographic correction process.Our study has some shortcomings. First, some errors remain in GlobeLand30 and national forest classification products compared with the actual surface cover types. Therefore, more detailed and accurate classification results need to be found as sample labels in the future. Second, the surface of sunny-shady slopes with different types of coverage is not considered in this study. Finally, U-Net should not be the sole focus. In subsequent research, we will attempt to select a variety of deep neural network classifiers to explore the classification accuracy changes before and after topographic correction of neural network models at a more precise scale to obtain further conclusions.

palabra clave

remote sensing classification;deep neural network;U-Net model;topographic correction;land cover classification;classification accuracy

References

  1. 1.
    Álvarez-Martínez J M, Silió-Calzada A and Barquín J. 2018. Can training data counteract topographic effects in supervised image classification? A sensitivity analysis in the Cantabrian Mountains (Spain). International Journal of Remote Sensing, 39(23): 8646-8669
  2. 2.
    Brown D G, Lusch D P and Duda K A. 1998. Supervised classification of types of glaciated landscapes using digital elevation data. Geomorphology, 21(3/4): 233-250
  3. 3.
    Chen C X, Hu C M, Huo L Z and Tang P. 2014. Effect of different radiation correction methods of Landsat TM data on land-cover remote sensing classification. Journal of Remote Sensing, 18(2): 320-334
  4. 4.
    Chen J, Chen J, Liao A P, Cao X, Chen L J, Chen X H, He C Y, Han G, Peng S, Lu M, Zhang W W, Tong X H and Mills J. 2015. Global land cover mapping at 30m resolution: a POK-based operational approach. ISPRS Journal of Photogrammetry and Remote Sensing, 103: 7-27
  5. 5.
    Chen J, Liao A P, Chen J, Peng S, Chen L J and Zhang H W. 2017. 30-meter global land cover data product- GlobeLand30. WorldGeomatics, 24(1): 1-8
  6. 6.
    Deng Y, Huang J and He M. 2018. Study on terrain radiometric correction of Landsat 8 in Chong Qing mountainous area. Rural Economy and Science-Technology, 29(11): 16-18
  7. 7.
    Gao Y N and Zhang W C. 2007. Variable empirical coefficient algorithm for removal of topographic effects on remotely sensed data from rugged terrain//Proceedings of 2007 IEEE International Geoscience and Remote Sensing Symposium. Barcelona: IEEE: 4733-4736
  8. 8.
    Gu D G and Gillespie A. 1998. Topographic normalization of Landsat tm images of forest based on subpixel sun-canopy-sensor geometry. Remote Sensing of Environment, 64(2): 166-175
  9. 9.
    Han X M, Li X H, Zhu R F and Shi X Y. 2010. Topographic correction method of remote sensing image considering Bi-directional reflectance//Proceedings of 2010 International Conference on Computer Application and System Modeling. Taiyuan, China: IEEE
  10. 10.
    He C, Chen J Z and Yue C R. 2014. Comparison test and research progress of normalization topographic correction model. Forest Inventory and Planning, 39(3): 12-16, 34
  11. 11.
    Isikdogan F, Bovik A C and Passalacqua P. 2017. Surface water mapping by deep learning. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 10(11): 4909-4918
  12. 12.
    Jia K, Li Q Z, Tian Y C and Wu B F. 2011. A review of classification methods of remote sensing imagery. Spectroscopy and Spectral Analysis, 31(10): 2618-2623
  13. 13.
    Jiang K, Hu C M, Yu K and Zhao Y C. 2014. Landsat TM/ETM+topographic correction method based on smoothed terrain and semi-empirical model. Journal of Remote Sensing, 18(2): 287-306
  14. 14.
    Kang X W and Feng Z K. 2011. An introduction to ASTER GDEM and procedure reading. Remote Sensing Information, (6): 69-72
  15. 15.
    Li C C, Wang J, Hu L Y, Yu L, Clinton N, Huang H B, Yang J and Gong P. 2014. A circa 2010 thirty meter resolution forest map for China. Remote Sensing, 6(6): 5325-5343
  16. 16.
    Li M Q, Im J and Beier C. 2013. Machine learning approaches for forest classification and change analysis using multi-temporal Landsat tm images over Huntington wildlife forest. GIScience and Remote Sensing, 50(4): 361-384
  17. 17.
    Li Y D, Hao Z B and Lei H. 2016. Survey of convolutional neural network. Journal of Computer Applications, 36(9): 2508-2515
  18. 18.
    Liu T and Chen X. 2018. Application of deep learning in globeland30-2010 product refinement. ISPRS-International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, XLII-3: 1111-1116
  19. 19.
    Ma Y F, Zhao J B, Li J and Qiao X S. 2018. Remote Sensing Image Classification Based on Artificial Neural Networks. Standardization of Surveying and Mapping, 34(4): 34-37
  20. 20.
    Maggiori E, Tarabalka Y, Charpiat G and Alliez P. 2016. Fully convolutional neural networks for remote sensing image classification//Proceedings of 2016 IEEE International Geoscience and Remote Sensing Symposium. Beijing, China: IEEE
  21. 21.
    Pimple U, Sitthi A, Simonetti D, Pungkul S, Leadprathom K and Chidthaisong A. 2017. Topographic correction of Landsat TM-5 and Landsat OLI-8 imagery to improve the performance of forest classification in the mountainous terrain of northeast Thailand. Sustainability, 9(2): 258
  22. 22.
    Ronneberger O, Fischer P and Brox T. 2015. U-Net: convolutional networks for biomedical image segmentation//Proceedings of the 18th International Conference on Medical Image Computing and Computer-Assisted Intervention. Munich: Springer: 234-241
  23. 23.
    Shelhamer E, Long J and Darrell T. 2017. Fully convolutional networks for semantic segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(4): 640-651
  24. 24.
    Soenen S A, Peddle D R and Coburn C A. 2005. SCS+C: A modified sun-canopy-sensor topographic correction in forested terrain. IEEE Transactions on Geoscience and Remote Sensing, 43(9): 2148-2159
  25. 25.
    Tang P, Zheng K, Shan X J, Hu C M, Huo L Z, Zhao L J and Li H Y. 2016. Framework of remote sensing image automatic processing with “invariant feature point set” as control data set. Journal of Remote Sensing, 20(5): 1126-1137
  26. 26.
    Vermote E, Justice C, Claverie M and Franch B. 2016. Preliminary analysis of the performance of the Landsat 8/OLI land surface reflectance product. Remote Sensing of Environment, 185: 46-56
  27. 27.
    Xu H M. 2018. Method research of high resolution remote sensing image classification based on U-net model of Deep learning. Chengdu: Southwest Jiaotong University

Leer el texto completo

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