Skip to main content
Springer Nature Link
Log in

Classify-normalize-classify

A novel data-driven framework for classifying forest pixels in remote sensing imagery

  • Research Article
  • Published:
Earth Science Informatics Aims and scope Submit manuscript

Abstract

Changing atmospheric conditions often result in a data distribution shift in remote sensing images for different dates and locations making it difficult to discriminate between various classes of interest. To alleviate this data shift issue, we introduce a novel supervised classification framework, called Classify-Normalize-Classify (CNC). The proposed scheme uses a two classifier approach where the first classifier performs a rough segmentation of the class of interest (COI) in the input image. Then, the median signal of the estimated COI regions is subtracted from all image pixels values to normalize them. Finally, the second classifier is applied to the normalized image to produce the refined COI segmentation. The proposed methodology was tested to detect deforestation using bitemporal Landsat 8 OLI images over the Amazon rainforest. The CNC framework compared favorably to benchmark masks of the PRODES program and state-of-the-art classifiers run on surface reflectance images provided by USGS.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Subscribe and save

Springer+ Basic
$34.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10

Similar content being viewed by others

Notes

  1. It is not required the top-of-atmosphere reflectance to be constant.

  2. Supposing that the erosion process of removing predicted forest pixels from the initial mask is unbiased.

  3. Despite the input images having 30 m resolution, all generated forest loss masks were evaluated at 60 m resolution by ignoring deforestation blobs smaller than 60 ×60 meters (see “Accuracy metrics”).

  4. All tiles used in our experiments are smaller than 1000 ×1000 pixels. For comparison, a typical Landsat image has dimensions of about 7750 ×7750 pixels.

  5. The App version of the surface reflectance images was “LaSRC_0.8.0”. In addition to Vermote et al. (2016), more information about the surface reflectance Landsat 8 OLI product can be found in https://landsat.usgs.gov/sites/default/files/documents/provisional_lasrc_product_guide_ee.pdf.

  6. This normalization should not be confused with the forest median centering normalization used in the CNC framework. The first uses statistics from all pixels from all tiles in the training set. The later uses only statistics from the forest pixels independently for each tile.

  7. Precision is also known in the remote sensing literature as the user’s accuracy for the positive class and sensitivity is also know as the producer’s accuracy for the positive class.

  8. These terms are also know as precision and sensitivity.

  9. All pixels in the Landsat image were processed including those located inside the no-data regions (black pixels). We could make it faster by avoiding processing those regions.

  10. The valid region of the image corresponds to an area of 185 km ×180 km.

References

  • Abadi M, Agarwal A, Barham P, Brevdo E, Chen Z, Citro C, Corrado GS, Davis A, Dean J, Devin M, Ghemawat S, Goodfellow I, Harp A, Irving G, Isard M, Jia Y, Jozefowicz R, Kaiser L, Kudlur M, Levenberg J, Mané D, Monga R, Moore S, Murray D, Olah C, Schuster M, Shlens J, Steiner B, Sutskever I, Talwar K, Tucker P, Vanhoucke V, Vasudevan V, Viégas F, Vinyals O, Warden P, Wattenberg M, Wicke M, Yu Y, Zheng X (2015) TensorFlow: Large-scale machine learning on heterogeneous systems. http://tensorflow.org/

  • Arango R B, Díaz I, Campos A, Canas E R, Combarro E F (2016) Automatic arable land detection with supervised machine learning. Earth Sci Inform 9(4):535–545

    Article  Google Scholar 

  • Bao N, Lechner A M, Fletcher A, Mellor A, Mulligan D, Bai Z (2012) Comparison of relative radiometric normalization methods using pseudo-invariant features for change detection studies in rural and urban landscapes. J Appl Remote Ses 6(1):063,578

    Google Scholar 

  • Berk A, Bernstein LS, Robertson DC (1987) Modtran: A moderate resolution model for lowtran. Tech. rep., DTIC Document. www.dtic.mil/docs/citations/ADA214337

  • Blaschke T (2010) Object based image analysis for remote sensing. ISPRS J Photogramm Remote Sens 65 (1):2–16

    Article  Google Scholar 

  • Canty M J, Nielsen A A (2008) Automatic radiometric normalization of multitemporal satellite imagery with the iteratively re-weighted mad transformation. Remote Sens Environ 112(3):1025–1036

    Article  Google Scholar 

  • Chollet F (2015) Keras. https://github.com/fchollet/keras

  • Dube T, Mutanga O (2015) Investigating the robustness of the new landsat-8 operational land imager derived texture metrics in estimating plantation forest aboveground biomass in resource constrained areas. ISPRS J Photogramm Remote Sens 108:12–32

    Article  Google Scholar 

  • Glorot X, Bengio Y (2010) Understanding the difficulty of training deep feedforward neural networks. In: AISTATS, vol 9, pp 249–256

  • Ham J, Lee DD, Saul LK (2005) Semisupervised alignment of manifolds. In: AISTATS, pp 120–127

  • Hansen M C, Roy D P, Lindquist E, Adusei B, Justice C O, Altstatt A (2008) A method for integrating MODIS and Landsat data for systematic monitoring of forest cover and change in the Congo Basin. Remote Sens Environ 112(5):2495–2513

    Article  Google Scholar 

  • Hansen M C, Potapov P V, Moore R, Hancher M, Turubanova S A, Tyukavina A, Thau D, Stehman S V, Goetz S J, Loveland T R, Kommareddy A, Egorov A, Chini L, Justice C O, Townshend J R G (2013) High-resolution global maps of 21st-century forest cover change. Science 342(2013):850–3

    Article  Google Scholar 

  • Hansen M C, Krylov A, Tyukavina A, Potapov P V, Turubanova S, Zutta B, Ifo S, Margono B, Stolle F, Moore R (2016) Humid tropical forest disturbance alerts using Landsat data. Environ Res Lett 11(3):034,008

    Article  Google Scholar 

  • Haralick R M, Shanmugam K, Dinstein I (1973) Textural features for image classification. IEEE Trans Syst, Man Cybern SMC 3(6):610–621

    Article  Google Scholar 

  • Hölbling D, Friedl B, Eisank C (2015) An object-based approach for semi-automated landslide change detection and attribution of changes to landslide classes in northern taiwan. Earth Sci Inform 8(2):327–335

    Article  Google Scholar 

  • Ioffe S, Szegedy C (2015) Batch normalization: Accelerating deep network training by reducing internal covariate shift. In: Proceedings International Conference Machine Learning, pp 448–456

  • Jr P S C (1988) An improved dark-object subtraction technique for atmospheric scattering correction of multispectral data. Remote Sens Environ 24(3):459–479

    Article  Google Scholar 

  • Maldonado MJDLR, Brown IF, Valeriano D, Duarte V (2007) Modificações no método do prodes para estimar a mudança da cobertura florestal na bacia trinacional do rio acre na região de fronteira entre bolívia, brasil e peru na amazonia sul-ocidental. In: Anais XIII Simpósio Brasileiro de Sensoriamento Remoto, pp 5903–5910

  • Marengo J A, Tomasella J, Soares W R, Alves L M, Nobre C A (2012) Extreme climatic events in the amazon basin. Theor Appl Climatol 107(1):73–85

    Article  Google Scholar 

  • Matasci G, Volpi M, Kanevski M, Bruzzone L, Tuia D (2015) Semisupervised transfer component analysis for domain adaptation in remote sensing image classification. IEEE Trans Geosci Remote Sens 53(7):3550–3564

    Article  Google Scholar 

  • Nair V, Hinton GE (2010) Rectified linear units improve restricted boltzmann machines. In: Proceedings International Conference Machine Learning, pp 807–814

  • Pan S J, Yang Q (2010) A survey on transfer learning. IEEE Trans Knowl Data Eng 22(10):1345–1359

    Article  Google Scholar 

  • Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V, Vanderplas J, Passos A, Cournapeau D, Brucher M, Perrot M, Duchesnay E (2011) Scikit-learn: Machine learning in Python. J Mach Learn Res 12:2825–2830

    Google Scholar 

  • Raval S, Shamsoddini A (2014) A monitoring framework for land use around kaolin mining areas through landsat tm images. Earth Sci Inform 7(3):153–163

    Article  Google Scholar 

  • Rossum G (1995) Python reference manual. Tech. Rep., Amsterdam. https://docs.python.org/2/reference/

    Google Scholar 

  • Roy D P, Wulder M, Loveland T, Woodcock C, Allen R, Anderson M, Helder D, Irons J, Johnson D, Kennedy R et al (2014) Landsat-8: Science and product vision for terrestrial global change research. Remote Sens Environ 145:154–172

    Article  Google Scholar 

  • Schott J R, Salvaggio C, Volchok W J (1988) Radiometric scene normalization using pseudoinvariant features. Remote Sens Environ 26(1):1–16

    Article  Google Scholar 

  • Shimabukuro Y E, dos Santos J R, Formaggio A R, Duarte V (2012) The brazilian amazon monitoring program: Prodes and deter projects. In: Rudorff BFT, Hansen MC (eds) Global forest monitoring from earth observation, CRC Press, Boca Raton, FL, chapter 9, pp 153–170

  • Singh A (1989) Digital change detection techniques using remotely-sensed data. Int J Remote Sens 10 (6):989–1003

    Article  Google Scholar 

  • Song C, Woodcock C E, Seto K C, Lenney M P, Macomber S A (2001) Classification and change detection using landsat TM data: When and how to correct atmospheric effects Remote Sens Environ 75(2):230–244

    Article  Google Scholar 

  • Srivastava N, Hinton G, Krizhevsky A, Sutskever I, Salakhutdinov R (2014) Dropout: A simple way to prevent neural networks from overfitting. J Mach Learn Res 15:1929–1958

    Google Scholar 

  • Tardy B, Rivalland V, Huc M, Hagolle O, Marcq S, Boulet G (2016) A software tool for atmospheric correction and surface temperature estimation of landsat infrared thermal data. Remote Sens 8(9):696

    Article  Google Scholar 

  • Tuia D, Camps-Valls G (2016) Kernel manifold alignment for domain adaptation. PloS one 11(2):1–25

    Article  Google Scholar 

  • Tuia D, Munoz-Mari J, Gomez-Chova L, Malo J (2013) Graph matching for adaptation in remote sensing. IEEE Trans Geosci Remote Sens 51(1):329–341

    Article  Google Scholar 

  • Tuia D, Persello C, Bruzzone L (2016) Domain adaptation for the classification of remote sensing data: An overview of recent advances. IEEE Trans Geosci Remote Sens 4(2):41–57

    Article  Google Scholar 

  • Vermote E, Justice C, Claverie M, Franch B (2016) Preliminary analysis of the performance of the landsat 8/oli land surface reflectance product. Remote Sens Environ 185:46–56

    Article  Google Scholar 

  • Vermote E F, Tanré D, Deuze J L, Herman M, Morcette J J (1997) Second simulation of the satellite signal in the solar spectrum, 6s: An overview. IEEE Trans Geosci Remote Sens 35(3):675–686

    Article  Google Scholar 

  • van der Walt S, Colbert S C, Varoquaux G (2011) The numpy array: A structure for efficient numerical computation. Comput Sci Eng 13(2):22–30

    Article  Google Scholar 

  • van der Walt S, Schönberger JL, Nunez-Iglesias J, Boulogne F, Warner JD, Yager N, Gouillart E, Yu T (2014) The scikit-image contributors scikit-image: image processing in Python. PeerJ 2:e453

Download references

Acknowledgments

This work was supported in part by the Brazilian National Council for Scientific and Technological Development (CNPq) [grant number 401113/2014-0].

Author information

Authors and Affiliations

  1. Federal University of Rio Grande do Sul, Institute of Informatics, Avenida Bento Gonçalves 9500, Porto Alegre, RS, 91501-970, Brazil

    Cesar Salgado, Maciel Zortea & Jacob Scharcanski

Authors
  1. Cesar Salgado
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Maciel Zortea
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Jacob Scharcanski
    View author publications

    You can also search for this author inPubMed Google Scholar

Corresponding author

Correspondence to Jacob Scharcanski.

Additional information

Communicated by: H. A. Babaie

Electronic supplementary material

Below is the link to the electronic supplementary material.

(PDF 75.0 KB)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Salgado, C., Zortea, M. & Scharcanski, J. Classify-normalize-classify. Earth Sci Inform 11, 77–97 (2018). https://doi.org/10.1007/s12145-017-0318-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s12145-017-0318-2

Keywords