Semisupervised Kernel Feature Extraction for Remote Sensing Image Analysis

This paper presents a novel semisupervised kernel partial least squares (KPLS) algorithm for nonlinear feature extraction to tackle both land-cover classification and biophysical parameter retrieval problems. The proposed method finds projections of the original input data that align with the target variable (labels) and incorporates the wealth of unlabeled information to deal with low-sized or underrepresented data sets. The method relies on combining two kernel functions: the standard radial-basis-function kernel based on labeled information and a generative, i.e., probabilistic, kernel directly learned by clustering the data many times and at different scales across the data manifold. The construction of the kernel is very simple and intuitive: Two samples should belong to the same class if they consistently belong to the same clusters at different scales. The effectiveness of the proposed method is successfully illustrated in multi- and hyperspectral remote sensing image classification and biophysical parameter estimation problems. Accuracy improvements in the range between +5% and 15% over standard principal component analysis (PCA), +4% and 15% over kernel PCA, and +3% and 10% over KPLS are obtained on several images. The average gain in the root-mean-square error of +5% and reductions in bias estimates of +3% are obtained for biophysical parameter retrieval compared to standard PCA feature extraction.

I. Introduction

FEATURE extraction has become an important topic in remote sensing data processing mainly due to the high dimensionality of data, as well as the high redundancy among spectral bands. The problem is ubiquitous and very common in remote sensing image analysis. Moreover, the high dimensionality of remote sensing data is often increased by stacking spatial, spectral, temporal, and multiangular features to the spectral channels for modeling additional information sources. Feature extraction consists of identifying the most discriminative variables for data classification or regression. These variables are often associated with the most relevant directions in the data distribution. For example, feature extraction is typically conducted for reducing the dimensionality of hyperspectral images and infrared sounder imagery before classification and parameter retrieval.