Visual Information Processing Group
Home Members Publications Projects Events Resources Ph. D.

Remote Sensing Image Classification with Large Scale Gaussian Processes


P. Morales, A. Pérez-Suay, R. Molina and G. Camps-Valls, “Remote Sensing Image Classification with Large Scale Gaussian Processes”, IEEE Transactions on Geoscience and Remote Sensing 2017. DOI: 10.1109/TGRS.2017.2758922.


Current remote sensing image classification problems have to deal with an unprecedented big amount of heterogeneous and complex data sources. Upcoming missions will soon provide large data streams that will make land cover/use classification difficult. Machine learning classifiers can help at this, and many methods are currently available. A popular kernel classifier is the Gaussian process classifier (GPC), since it approaches the classification problem with a solid probabilistic treatment, thus yielding confidence intervals for the predictions as well as very competitive results to state-of-the-art neural networks and support vector machines. However, its computational cost is prohibitive for large scale applications, and constitutes the main obstacle precluding wide adoption. This paper tackles this problem by introducing two novel efficient methodologies for GP classification. We first include the standard random Fourier features approximation into GPC, which largely decreases its computational cost and permits large scale remote sensing image classification. In addition, we propose a model which avoids randomly sampling a number of Fourier frequencies, and alternatively {\em learns} the optimal ones within a variational Bayes approach. The performance of the proposed methods is illustrated in complex problems of cloud detection from multispectral imagery and infrared sounding data. Excellent empirical results support the proposal in both computational cost and accuracy.

Classification maps video

This video shows six different figures. The "RGB", "infrared band", "L2 cloud mask", and "prediction" ones correspond with those included in Figure 6 of the submitted manuscript. "Land cover" shows the water (blue) or land (red) cover of the Dakhla landmark (and thus it does not change during the video). The "cumulative chip accuracy" includes the accuracy of the method on each video frame (blue line) as well as the cumulative accuracy (red line).



The MATLAB code of the proposed method can be downloaded here.

Also available in GitHub:

Visual Image Processing
University of Granada