Classification of PolSAR Images by Stacked Random Forests

This paper proposes the use of Stacked Random Forests (SRF) for the classification of Polarimetric Synthetic Aperture Radar images. SRF apply several Random Forest instances in a sequence where each individual uses the class estimate of its predecessor as an additional feature. To this aim, the internal node tests are designed to work not only directly on the complex-valued image data, but also on spatially varying probability distributions and thus allow a seamless integration of RFs within the stacking framework. Experimental results show that the classification performance is consistently improved by the proposed approach, i.e., the achieved accuracy is increased by 4 % and 7 % for one fully- and one dual-polarimetric dataset. This increase only comes at the cost of a linear increased training and prediction time, which is rather limited as the method converges quickly.

[1]  David H. Wolpert,et al.  Stacked generalization , 1992, Neural Networks.

[2]  Olaf Hellwich,et al.  Random Forests for building detection in polarimetric SAR data , 2010, 2010 IEEE International Geoscience and Remote Sensing Symposium.

[3]  Olaf Hellwich,et al.  Skipping the real world: Classification of PolSAR images without explicit feature extraction , 2017 .

[4]  Mingsheng Liao,et al.  Nonlinear Compressed Sensing-Based LDA Topic Model for Polarimetric SAR Image Classification , 2014, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[5]  Thomas L. Ainsworth,et al.  Unsupervised classification of polarimetric synthetic aperture Radar images using fuzzy clustering and EM clustering , 2005, IEEE Transactions on Geoscience and Remote Sensing.

[6]  Haipeng Wang,et al.  Complex-Valued Convolutional Neural Network and Its Application in Polarimetric SAR Image Classification , 2017, IEEE Transactions on Geoscience and Remote Sensing.

[7]  Tin Kam Ho,et al.  The Random Subspace Method for Constructing Decision Forests , 1998, IEEE Trans. Pattern Anal. Mach. Intell..

[8]  Knut Conradsen,et al.  A test statistic in the complex Wishart distribution and its application to change detection in polarimetric SAR data , 2003, IEEE Trans. Geosci. Remote. Sens..

[9]  N. Ayache,et al.  Log‐Euclidean metrics for fast and simple calculus on diffusion tensors , 2006, Magnetic resonance in medicine.

[10]  Olaf Hellwich,et al.  Evaluation of tree creation methods within random forests for classification of PolSAR images , 2015, 2015 IEEE International Geoscience and Remote Sensing Symposium (IGARSS).

[11]  Mandy Eberhart,et al.  Decision Forests For Computer Vision And Medical Image Analysis , 2016 .

[12]  Mariana Belgiu,et al.  Random forest in remote sensing: A review of applications and future directions , 2016 .

[13]  Gabriele Moser,et al.  Supervised High-Resolution Dual-Polarization SAR Image Classification by Finite Mixtures and Copulas , 2011, IEEE Journal of Selected Topics in Signal Processing.

[14]  Leo Breiman,et al.  Stacked regressions , 2004, Machine Learning.

[15]  Robert Jenssen,et al.  Spectral Clustering of Polarimetric SAR Data With Wishart-Derived Distance Measures , 2007 .

[16]  Jocelyn Chanussot,et al.  A novel approach to polarimetric SAR data processing based on Nonlinear PCA , 2014, Pattern Recognit..

[17]  Eric Pottier,et al.  A review of target decomposition theorems in radar polarimetry , 1996, IEEE Trans. Geosci. Remote. Sens..

[18]  E. Pottier,et al.  Polarimetric Radar Imaging: From Basics to Applications , 2009 .

[19]  Leo Breiman,et al.  Bagging Predictors , 1996, Machine Learning.

[20]  Vincent Lepetit,et al.  Keypoint recognition using randomized trees , 2006, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[21]  Ronny Hänsch,et al.  Complex-Valued Multi-Layer Perceptrons -An Application to Polarimetric SAR Data , 2010 .

[22]  Piotr Tokarczyk,et al.  Features, Color Spaces, and Boosting: New Insights on Semantic Classification of Remote Sensing Images , 2015, IEEE Transactions on Geoscience and Remote Sensing.

[23]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[24]  N. R. Goodman.  Statistical analysis based on a certain multivariate complex Gaussian distribution , 1963 .

[25]  Ron Kwok,et al.  Classification of multi-look polarimetric SAR imagery based on complex Wishart distribution , 1994 .

[26]  Feng Zhou,et al.  Tensorial Independent Component Analysis-Based Feature Extraction for Polarimetric SAR Data Classification , 2015, IEEE Transactions on Geoscience and Remote Sensing.

[27]  Ronny Hänsch.  Generic object categorization in PolSAR images - and beyond , 2014 .

[28]  Yu Zhou,et al.  Polarimetric SAR Image Classification Using Deep Convolutional Neural Networks , 2016, IEEE Geoscience and Remote Sensing Letters.

[29]  Lorenzo Bruzzone,et al.  An advanced system for the automatic classification of multitemporal SAR images , 2004, IEEE Transactions on Geoscience and Remote Sensing.

[30]  Gabriele Moser,et al.  Partially Supervised classification of remote sensing images through SVM-based probability density estimation , 2005, IEEE Transactions on Geoscience and Remote Sensing.

[31]  Mingsheng Liao,et al.  Texture Classification of PolSAR Data Based on Sparse Coding of Wavelet Polarization Textons , 2013, IEEE Transactions on Geoscience and Remote Sensing.

[32]  Henri Maître,et al.  A new statistical model for Markovian classification of urban areas in high-resolution SAR images , 2004, IEEE Transactions on Geoscience and Remote Sensing.

[33]  Olaf Hellwich,et al.  Complex-Valued Convolutional Neural Networks for Object Detection in PolSAR data , 2010 .

[34]  Joachim Denzler,et al.  Semantic Segmentation with Millions of Features: Integrating Multiple Cues in a Combined Random Forest Approach , 2012, ACCV.

[35]  Frédéric Barbaresco,et al.  Interactions between Symmetric Cone and Information Geometries: Bruhat-Tits and Siegel Spaces Models for High Resolution Autoregressive Doppler Imagery , 2009, ETVC.

[36]  Jean-Marie Nicolas,et al.  Statistical models for SAR amplitude data: A unified vision through Mellin transform and Meijer functions , 2016, 2016 24th European Signal Processing Conference (EUSIPCO).

[37]  Gabriele Moser,et al.  Kernel-based classification in complex-valued feature spaces for polarimetric SAR data , 2014, 2014 IEEE Geoscience and Remote Sensing Symposium.