Review of machine learning methods for Big satellite Data classification
DOI:
https://doi.org/10.20535/SRIT.2308-8893.2018.1.04Keywords:
machine learning, deep learning, convolutional neural network, big data classificationAbstract
With the appearance of free access to Big satellite data, the development of machine learning methods based on geospatial data, in particular satellite data, is becoming more and more relevant. In this paper, we consider and analyze the peculiarities of the basic machine learning methods and results of their application to the tasks of land cover classification based on high resolution satellite data. Special attention is paid to deep architectures, in particular, convolutional neural networks, which nowadays are the most powerful and precise method for visual pattern recognizing. We determine the main advantages of the deep learning methods over the traditional approaches to the classification tasks, that have been used over the last decades and based on expert knowledge to the features extraction from the input data.References
Bishop C.M. Pattern Recognition and Machine Learning / C.M. Bishop. — NY: Springer. — 2006.
Zgurovsky M.Z. System Analysis: Theory and Applications / M.Z. Zgurovsky, N.D. Pankratova // Springer. — 2007. — 448 p.
Novikov A. The synthesis of information protection systems with optimal properties / A. Novikov, A. Rodionov // Complexity and Security. — Vol. 37. — 2008. — 307 p.
LeCun Y. Deep learning / Y. LeCun, B. Yoshua, H. Geoffrey // Nature. — Vol. 521, N 7553. — 2015. — P. 436–444.
Ghamisi P. Advanced Spectral Classifiers for Hyperspectral Images: A review / P. Ghamisi, J. Plaza, Y. Chen et al // IEEE Geoscience and Remote Sensing Magazine. — Vol. 5, N 1. — 2017. — P. 8–32.
Fukunaga K. Introduction to statistical pattern recognition / K. Fukunaga // Academic press, 2013. — 591 p.
Maulik U. Remote Sensing Image Classification: A survey of support-vector-machine-based advanced techniques / U. Maulik, D. Chakraborty // IEEE Geoscience and Remote Sensing Magazine. — Vol. 5, N 1. — 2017. — P. 33–52.
Gislason P.O. Random forests for land cover classification / P.O. Gislason, J.A. Benediktsson, J.R. Sveinsson // Pattern Recognition Letters. — Vol. 27, N 4. — 2006. — P. 294–300.
McCulloch W.S. A logical calculus of the ideas immanent in nervous activity / W.S. McCulloch, W. Pitts // The bulletin of mathematical biophysics. — Vol. 5, N 4. — 1943. — P. 115–133.
Rumelhart D.E. Learning representations by back-propagating errors / D.E. Rumelhart, G.E. Hinton, R.J. Williams // Cognitive modeling. — Vol. 5, N 3. — 1988. — P. 213–220.
Amari S.I. Statistical theory of learning curves under entropic loss criterion / S.I. Amari, N. Murata // Neural Computation. — Vol. 5, N 1. — 1993. — P. 140–153.
Schmidhuber J. Deep learning in neural networks: An overview / J. Schmidhuber // Neural networks. — Vol. 61. — 2015. — P. 85–117.
Rumelhart D.E. Learning internal representations by error propagation / D.E. Rumelhart, G.E. Hinton, R.J. Williams // Parallel distributed processing. MIT Press. — 1986. — P. 318–362.
Duchi J. Adaptive subgradient methods for online learning and stochastic optimization / J. Duchi, E. Hazan, Y. Singer // Journal of Machine Learning Research. — 2011. — P. 2121–2159.
Patel V. Kalman-Based Stochastic Gradient Method with Stop Condition and Insensitivity to Conditioning / V. Patel // SIAM Journal on Optimization. — Vol. 26, N 4. — 2016. — P. 2620–2648.
Kingma D. Adam: A method for stochastic optimization / D. Kingma, J. Ba // arXiv preprint arXiv. — Vol. 1412, N 6980. — 2014. — P. 1–15.
Haykin S. Neural networks and learning machines / S. Haykin // Upper Saddle River. — NJ, USA: Pearson, 2009. — Vol. 3. — 938 p.
Hornik K. Multilayer feedforward networks are universal approximators / K. Hornik, M. Stinchcombe, H. White // Neural networks. — Vol. 2, N 5. — 1989. — P. 359–366.
Liu P. SVM or deep learning? A comparative study on remote sensing image classification / P. Liu, K.K.R. Choo, L. Wang, F. Huang // Soft Computing. — Vol. 21, N 23. — 2017. — P. 7053–7065.
Huang F.J. Large-scale learning with SVM and convolutional nets for generic object categorization / F.J. Huang, Y. LeCun // IEEE Computer Society Conference on Computer Vision and Pattern Recognition. — 2006. — P. 284–291. DOI: 10.1109/CVPR.2006.164.
Pirotti F. Benchmark of machine learning methods for classification of a Sentinel-2 image / F. Pirotti, F. Sunar, M. Piragnolo // International Archives of the Photogrammetry, Remote Sensing & Spatial Information Sciences. — Vol. 41. — 2016. — P. 335–340.
Breiman L. Random forests / L. Breiman // Machine learning. — Vol. 45, N 1. — 2001. — P. 5–32.
Ivakhnenko A.G. The group method of data handling-a rival of the method of stochastic approximation / A.G. Ivakhnenko // Soviet Automatic Control. — Vol. 13, N 3. — 1968. — P. 43–55.
Ivakhnenko A.G. Polynomial theory of complex systems / A.G. Ivakhnenko // IEEE transactions on Systems, Man, and Cybernetics. — Vol. 1, N 4. — 1971. — P. 364–378.
Fukushima K. Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position / K. Fukushima // Biol. Cybern. — Vol. 36. —1980. — P. 193–202. Doi:10.1007/bf00344251.
LeCun Y. Backpropagation applied to handwritten zip code recognition / Y. LeCun, B. Boser, J.S. Denker et al. // Neural computation. — Vol. 1, N 4. — 1989. — P. 541–551.
Hinton G.E. A fast learning algorithm for deep belief nets / G.E. Hinton, S. Osindero, Y.W. Teh // Neural computation. — Vol. 18, N 7. — 2006. — 1527–1554 p.
Connor J.T. Recurrent neural networks and robust time series prediction / J.T. Connor, R.D. Martin, L.E. Atlas // IEEE transactions on neural networks. — Vol. 5, N 2. — 1994. — P. 240–254.
Hochreiter S. Untersuchungen zu dynamischen neuronalen Netzen / S. Hochreiter // Diss. diploma thesis, institut für informatik, lehrstuhl prof. brauer, technische universität münchen. — 1991.
Hochreiter S. Long short-term memory / S. Hochreiter, J. Schmidhuber // Neural computation. — Vol. 9, N 8. — 1997. — P. 1735–1780.
Ding J. Convolutional neural network with data augmentation for SAR target recognition / J. Ding, B. Chen, H. Liu, M. Huang // IEEE Geoscience and Remote Sensing Letters. — Vol. 13, N 3. — 2016. — P. 364–368.
Krizhevsky A. Imagenet classification with deep convolutional neural networks / A. Krizhevsky, I. Sutskever, G.E. Hinton // In Advances in neural information processing systems. — 2012. —P. 1097–1105.
Ngiam J. Multimodal deep learning / J. Ngiam, A. Khosla, M. Kim et al. // In Proceedings of the 28th international conference on machine learning (ICML-11). — 2011. — P. 689–696.
Chen X. Vehicle detection in satellite images by hybrid deep convolutional neural networks / X. Chen, S. Xiang, C.L. Liu, C.H. Pan // IEEE Geoscience and remote sensing letters. — Vol. 11, N 10. — 2014. — P. 1797–1801.
Hinton G. Deep neural networks for acoustic modeling in speech recognition: The shared views of four research groups / G. Hinton, L. Deng, D. Yu et al. // IEEE Signal Processing Magazine. — Vol. 29, N 6. — 2012. — P. 82–97.
Helmstaedter M. Connectomic reconstruction of the inner plexiform layer in the mouse retina / M. Helmstaedter, K.L. Briggman, S.C. Turaga et al. // Nature. — Vol. 500, N 7461. — 2013. — P. 168–174.
Sutskever I. Sequence to sequence learning with neural networks / I. Sutskever, O. Vinyals, Q.V. Le // In Advances in neural information processing systems. — 2014. — P. 3104–3112.
Drusch M. Sentinel-2: ESA's optical high-resolution mission for GMES operational services / M. Drusch, U. Del Bello, S. Carlier et al. // Remote Sensing of Environment. — Vol. 120. — 2012. — P. 25–36.
Torres R. GMES Sentinel-1 mission / R. Torres, P. Snoeij, D. Geudtner et al. // Remote Sensing of Environment. — Vol. 120. — 2012. — P. 9–24.
Roy D.P. Landsat-8: Science and product vision for terrestrial global change research / D.P. Roy, M.A. Wulder, T.R. Loveland et al. // Remote Sensing of Environment. — Vol. 145. — 2014. — P. 154–172.
Zhang J. Multi-source remote sensing data fusion: status and trends / J. Zhang // International Journal of Image and Data Fusion. — Vol. 1, N 1. — 2010. — P. 5–24.
Dalla M.M. Challenges and opportunities of multimodality and data fusion in remote sensing / M.M. Dalla, S. Prasad, F. Pacifici et al. // Proceedings of the IEEE. — Vol. 103, N 9. — 2015. — P. 1585–1601.
Gómez-Chova L. Multimodal classification of remote sensing images: A review and future directions / L. Gómez-Chova, D. Tuia, G. Moser, G. Camps-Valls // Proceedings of the IEEE. – Vol. 103, N 9. — 2015. — 1560–1584.
Kuwata K. Estimating crop yields with deep learning and remotely sensed data / K. Kuwata, R. Shibasaki // Geoscience and Remote Sensing Symposium (IGARSS). — 2015. — P. 858–861.
McNairn H. Early season monitoring of corn and soybeans with TerraSAR-X and RADARSAT-2 / H. McNairn, A. Kross, D. Lapen, R. Caves, J. Shang // International Journal of Applied Earth Observation and Geoinformation. — Vol. 28. — 2014. — P. 252–259.
Khatami R. A meta-analysis of remote sensing research on supervised pixel-based land-cover image classification processes: General guidelines for practitioners and future research / R. Khatami, G. Mountrakis, S.V. Stehman // Remote Sensing of Environment. — Vol. 177. — 2016. — P. 89–100.
Abdiansah A. Time Complexity Analysis of Support Vector Machines (SVM) in LibSVM / A. Abdiansah, R. Wardoyo // International Journal of Computer Applications. — Vol. 128, N 3. — 2015. — P. 28–34.
Chapelle O. Training a Support Vector Machine in the Primal / O. Chapelle // Neural Computation. — Vol. 19, N 5. — 2007. — P. 1155–1178.
Sun X. meProp: Sparsified back propagation for accelerated deep learning with reduced overfitting / X. Sun, X. Ren, S. Ma, H. Wang // arXiv preprint arXiv. — Vol. 1706, N 06197. — 2017. — P. 1–10.
Gallagher M. Multi-layer perceptron error surfaces: visualization, structure and mod-elling / M. Gallagher // Doctoral dissertation. — 2000.
Han M. (2012). Remote sensing image classification based on neural network ensemble algorithm / M. Han, X. Zhu, W.Yao // Neurocomputing. — Vol. 78, N 1. — 2012. — P. 133–138.
Huang X. An SVM ensemble approach combining spectral, structural, and semantic features for the classification of high-resolution remotely sensed imagery / X. Huang, L. Zhang // IEEE Transactions on Geoscience and Remote Sensing. — Vol. 51, N 1. — 2013. — P. 257–272.
Lavreniuk M.S. Large-scale classification of land cover using retrospective satellite data / M.S. Lavreniuk, S.V. Skakun, A.J. Shelestov et al. // Cybernetics and Systems Analysis. — Vol. 52, N 1. — 2016. — P. 127–138.
Kussul N. Parcel-based crop classification in ukraine using landsat-8 data and sentinel-1A data / N. Kussul, G. Lemoine, F.J. Gallego et al. // IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing. — Vol. 9, N 6. — 2016. — P. 2500–2508.
Chen Y. Deep learning-based classification of hyperspectral data / Y. Chen, Z. Lin, X. Zhao et al. // IEEE Journal of Selected topics in applied earth observations and remote sensing. — Vol. 7, N 6. — 2014. — P. 2094–2107.
Zhao W. Learning multiscale and deep representations for classifying remotely sensed imagery / W. Zhao, S. Du // ISPRS Journal of Photogrammetry and Remote Sensing. — Vol. 113. — 2016. — P. 155–165.
Kussul N.N. Land cover changes analysis based on deep machine learning technique / N.N. Kussul, N.S. Lavreniuk, A.Y. Shelestov et al. // Journal of Automation and Information Sciences. — Vol. 48, N 5. — 2016. — P. 42–54.
Kussul N. Geospatial Intelligence and Data Fusion Techniques for Sustainable Development Problems / N. Kussul, A. Shelestov, R. Basarab et al. // ICTERI. — 2015. — P. 196–203.
Zhang F. Saliency-guided unsupervised feature learning for scene classification / F. Zhang, B. Du, L. Zhang // IEEE Transactions on Geoscience and Remote Sensing. — Vol. 53, N 4. — 2015. — P. 2175–2184.
Zhang F. Scene classification via a gradient boosting random convolutional network framework / F. Zhang, B. Du, L. Zhang // IEEE Transactions on Geoscience and Remote Sensing. — Vol. 54, N 3. — 2016. — P. 1793–1802.
Lary D.J. Machine learning in geosciences and remote sensing / D.J. Lary, A.H. Alavi, A.H. Gandomi, A.L. Walker // Geoscience Frontiers. — Vol. 7, N 1.— 2016. — P. 3–10.
Peña-Barragán J.M. Object-based crop identification using multiple vegetation indices, textural features and crop phenology / J.M. Peña-Barragán, M.K. Ngugi, R.E. Plant, J. Six // Remote Sensing of Environment. — Vol. 115, N 6. — 2011. — P. 1301–1316.
Mnih V. Learning to detect roads in high-resolution aerial images / V. Mnih, G. Hinton // Computer Vision–ECCV. — 2010. — P. 210–223.
Mnih V. Learning to label aerial images from noisy data / V. Mnih, G. Hinton // Proceedings of the 29th International Conference on Machine Learning (ICML-12). — 2012. — P. 1–8.
Geng J. High-resolution SAR image classification via deep convolutional autoencoders / J. Geng, J. Fan, H. Wang et al. // IEEE Geoscience and Remote Sensing Letters. — Vol. 12, N 11. — 2015. — P. 2351–2355.
Chen Y. Spectral–spatial classification of hyperspectral data based on deep belief network / Y. Chen, X. Zhao, X. Jia // IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing. — Vol. 8, N 6. — 2015. — P. 2381–2392.
Liang H. Hyperspectral imagery classification using sparse representations of convolutional neural network features / H. Liang, Q. Li // Remote Sensing. — Vol. 8, N 2. — 2016. — P. 1–16.
Lyu H. Learning a transferable change rule from a recurrent neural network for land cover change detection / H. Lyu, H. Lu, L. Mou // Remote Sensing. — Vol. 8, N 6. — 2016. — P. 1–22.
Ishii T. Surface object recognition with CNN and SVM in Landsat 8 images / T. Ishii, R. Nakamura, H. Nakada et al. // In Machine Vision Applications (MVA), 14th IAPR International Conference. — 2015. — P. 341–344.
Deng L. Deep learning: methods and applications / L. Deng, D. Yu // Foundations and Trends® in Signal Processing. — Vol. 7, N 3–4. — 2014. — P. 197–387.
Längkvist M. Classification and segmentation of satellite orthoimagery using convolutional neural networks / M. Längkvist, A. Kiselev, M. Alirezaie, A. Loutfi // Remote Sensing. — Vol. 8, N 4. — 2016. — P. 1–21.
Dauphin Y.N. Identifying and attacking the saddle point problem in high-dimensional non-convex optimization / Y.N. Dauphin, R. Pascanu, C. Gulcehre et al. // In Advances in neural information processing systems. — 2014. — P. 2933–2941.
Choromanska A. The loss surfaces of multilayer networks / A. Choromanska, M. Henaff, M. Mathieu et al. // In Artificial Intelligence and Statistics. — 2015. — P. 192–204.
Hu W. Deep convolutional neural networks for hyperspectral image classification / W. Hu, Y. Huang, L. Wei et al. // Journal of Sensors. — 2015. — P. 1–13.
