Biophysical parameters estimation from remotely sensed images by a multiple criteria active learning method

This paper presents a novel multıple criteria active learning method developed in the framework of ε-insensitive support vector regression (SVR) to estimate biophysical parameters from remotely sensed images. The proposed active learning method chooses the most informative and representative unlabeled samples by jointly evaluating three criteria: i) relevancy, ii) diversity, and iii) density of samples. All three criteria are implemented according to the SVR properties and are applied in two clustering-based successive steps. In the first step, a novel measure to select the most relevant samples that have high probability to be located either outside or on the boundary of the ε-tube of SVR is defined. In the second step, a novel measure to select diverse samples among the relevant patterns from the high density regions in the feature space is defined to better model the SVR learning function. Experiments applied to the estimation of tree stem volume show the effectiveness of the proposed method.