Abstract
This research is motivated by sustainability problems of oil palm expansion. Fast-growing industrial Oil Palm Plantations (OPPs) in the tropical belt of Africa, Southeast Asia and parts of Brazil lead to significant loss of rainforest and contribute to the global warming by the corresponding decrease of carbon dioxide absorption. We propose a novel approach to monitoring of the expansion of OPPs based on an application of state-of-the-art Fully Convolutional Neural Networks (FCNs) to solve Semantic Segmentation Problem for Landsat imagery. The proposed approach significantly outperforms per-pixel classification methods based on Random Forest using texture features, NDVI, and all Landsat bands. Moreover, the trained FCN is robust to spatial and temporal shifts of input data. The paper provides a proof of concept that FCNs as semi-automated methods enable OPPs mapping of entire countries and may serve for yearly detection of oil palm expansion.
This research was supported by Russian Science Foundation, grant no. 14-11-00109.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Object-oriented classifiers first perform image segmentation to merge pixels into objects and then apply classification based on the objects [19] (but not on an individual pixel as per-pixel classifiers do). Spectral-Color algorithms are trained using multi-spectral features of each pixel.
- 2.
In GEE, rectangles are represented by the minimum and maximum corners as lists (xMin, yMin, xMax, yMax). The ROI is defined by \((110.3, -3, 113.6, -1.5)\).
- 3.
An image from Landsat 8 has 11 bands, ‘B1’ – ‘B11’. NDVI stands for Normalized Difference Vegetation Index that varies between \(-1\) and 1 and utilizes the relationship between vegetation brightness in the red and infrared bands. Namely, absorption of red light (‘B4’) and reflection of infrared radiation (‘B5’) by plants allows to separate non-vegetation (e.g., water, bare soil, man-made objects) from living vegetation. Basically, NDVI serves as an index of photosynthesis in a pixel.
- 4.
Implemented on the basis of https://github.com/shelhamer/fcn.berkeleyvision.org.
- 5.
Caffe deep learning framework http://caffe.berkeleyvision.organized.
- 6.
For models with features ‘G(10)’ included, we sampled 20000 random pixels due to computational constraints of GEE.
- 7.
The training was performed in R using library ‘randomForest’ with all other default parameters except number of trees to grow was set to 300.
- 8.
The image size is \(2048 \times 1688\) pixels; Lat.: 1.14868; Long.: 99.18078. The distance to the boundary of Kalimantan ROI is around 1300Â km.
- 9.
The image size is \(1336 \times 1069\) pixels; Lat.: \(-1.93356\); Long.: 111.21357.
References
PASCAL VOC 2012. www.host.robots.ox.ac.uk/pascal/VOC/voc2012/
Austin, K., Mosnier, A., Pirker, J., McCallum, I., Fritz, S., Kasibhatla, P.: Shifting patterns of oil palm driven deforestation in Indonesia and implications for zero-deforestation commitments. Land Use Policy 69, 41–48 (2017). https://doi.org/10.1016/j.landusepol.2017.08.036. http://www.sciencedirect.com/science/article/pii/S0264837717301552
Belgiu, M., Drãguţ, L.: Random forest in remote sensing: a review of applications and future directions. ISPRS J. Photogram. Remote Sens. 114, 24–31 (2016). https://doi.org/10.1016/j.isprsjprs.2016.01.011. http://www.sciencedirect.com/science/article/pii/S0924271616000265
Carlson, K.M., Curran, L.M., Asner, G.P., Pittman, A.M., Trigg, S.N., Marion Adeney, J.: Carbon emissions from forest conversion by Kalimantan oil palm plantations. Nat. Clim. Change 3, 283–287 (2013). https://doi.org/10.1038/nclimate1702
Chen, L., Kokkinos, I., Murphy, K., Yuille, A.: DeepLab: semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs. IEEE Trans. Pattern Anal. Mach. Intell. 40, 834–848 (2018). https://doi.org/10.1109/TPAMI.2017.2699184
Chong, K.L., Kanniah, K.D., Pohl, C., Tan, K.P.: A review of remote sensing applications for oil palm studies. Geo-spatial Inf. Sci. 20(2), 184–200 (2017). https://doi.org/10.1080/10095020.2017.1337317
Conners, R.W., Trivedi, M.M., Harlow, C.A.: Segmentation of a high-resolution urban scene using texture operators. Comput. Vis. Graph. Image Process. 25(3), 273–310 (1984). https://doi.org/10.1016/0734-189X(84)90197-X
Daliman, S., Rahman, S., Bakar, S., Busu, I.: Segmentation of oil palm area based on GLCM-SVM and NDVI. In: IEEE Region 10 Symposium, pp. 645–650 (2014). https://doi.org/10.1109/TENCONSpring.2014.6863113
Fu, G., Liu, C., Zhou, R., Sun, T., Zhang, Q.: Classification for high resolution remote sensing imagery using a fully convolutional network. Remote Sens. 9(5), 498 (2017). https://doi.org/10.3390/rs9050498. http://www.mdpi.com/2072-4292/9/5/498
Garcia-Garcia, A., Orts-Escolano, S., Oprea, S., Villena-Martinez, V., Jose Garcia-Rodriguez, V.: A review on deep learning techniques applied to semantic segmentation. Manuscript 1 (2017)
Gaveau, D.L.A., et al.: Rapid conversions and avoided deforestation: examining four decades of industrial plantation expansion in Borneo. Sci. Rep. 6 (2016). https://doi.org/10.1038/srep32017
Glorot, X., Bengio, Y.: Understanding the difficulty of training deep feedforward neural networks. In: Teh, Y.W., Titterington, M. (eds.) Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, Proceedings of Machine Learning Research, vol. 9, pp. 249–256. PMLR, Sardinia, 13–15 May 2010. http://proceedings.mlr.press/v9/glorot10a.html
Gutiérrez-Vélez, V.H., DeFries, R.: Annual multi-resolution detection of land cover conversion to oil palm in the Peruvian Amazon. Remote Sens. Environ. 129, 154–167 (2013). https://doi.org/10.1016/j.rse.2012.10.033. http://www.sciencedirect.com/science/article/pii/S003442571200421X
Hansen, M.C., et al.: High-resolution global maps of 21st-century forest cover change. Science 342(6160), 850–853 (2013). https://doi.org/10.1126/science.1244693. http://science.sciencemag.org/content/342/6160/850
Haralick, R., Shanmugam, K., Dinstein, I.: Textural features for image classification. IEEE Trans. Syst. Man Cybern. SMC–3, 610–621 (1973). https://doi.org/10.1109/TSMC.1973.4309314
Huang, Z., Pan, Z., Lei, B.: Transfer learning with deep convolutional neural network for SAR target classification with limited labeled data. Remote Sens. 9(9), 907 (2017). https://doi.org/10.3390/rs9090907. http://www.mdpi.com/2072-4292/9/9/907
Kamiran, N., Sarker, M.L.R.: Exploring the potential of high resolution remote sensing data for mapping vegetation and the age groups of oil palm plantation. IOP Conf. Ser.: Earth Environ. Sci. 18(1), 012181 (2014). http://stacks.iop.org/1755-1315/18/i=1/a=012181
Lee, J.S.H., Wich, S., Widayati, A., Koh, L.P.: Detecting industrial oil palm plantations on Landsat images with Google Earth Engine. Remote Sens. Appl. Soc. Environ. 4, 219–224 (2016). https://doi.org/10.1016/j.rsase.2016.11.003. https://www.sciencedirect.com/science/article/pii/S235293851630129X
Lu, D., Weng, Q.: A survey of image classification methods and techniques for improving classification performance. Int. J. Remote Sens. 28(5), 823–870 (2007). https://doi.org/10.1080/01431160600746456
Mucherino, A., Papajorgji, P.J., Pardalos, P.M.: Data Mining in Agriculture, 1st edn. Springer, New York (2009). https://doi.org/10.1007/978-0-387-88615-2
Nooni, I., Duker, A., Van Duren, I., Addae-Wireko, L., Osei Jnr, E.: Support vector machine to map oil palm in a heterogeneous environment. Int. J. Remote Sens. 35(13), 4778–4794 (2014). https://doi.org/10.1080/01431161.2014.930201
Petersen, R., et al.: Mapping tree plantations with multispectral imagery: preliminary results for seven tropical countries. Technical report, World Resources Institute (2016). www.wri.org/publication/mapping-tree-plantations
Shelhamer, E., Long, J., Darrell, T.: Fully convolutional networks for semantic segmentation. IEEE Trans. Pattern Anal. Mach. Intell. 39, 640–651 (2017). https://doi.org/10.1109/TPAMI.2016.2572683
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2018 Springer Nature Switzerland AG
About this paper
Cite this paper
Baklanov, A., Khachay, M., Pasynkov, M. (2018). Application of Fully Convolutional Neural Networks to Mapping Industrial Oil Palm Plantations. In: van der Aalst, W., et al. Analysis of Images, Social Networks and Texts. AIST 2018. Lecture Notes in Computer Science(), vol 11179. Springer, Cham. https://doi.org/10.1007/978-3-030-11027-7_16
Download citation
DOI: https://doi.org/10.1007/978-3-030-11027-7_16
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-11026-0
Online ISBN: 978-3-030-11027-7
eBook Packages: Computer ScienceComputer Science (R0)