Superpixel-Based Active Learning and Online Feature Importance Learning for Hyperspectral Image Analysis

The rapid development of multichannel optical imaging sensors has led to increased utilization of hyperspectral data for remote sensing. For classification of hyperspectral data, an informative training set is necessary for ensuring robust performance. However, in remote sensing and other image analysis applications, labeled samples are often difficult, expensive, and time-consuming to obtain. This makes active learning (AL) an important part of an image analysis framework-AL aims to efficiently build a representative and efficient library of training samples that are most informative for the underlying classification task. This paper proposes an AL framework that leverages from superpixels. A spatial-spectral AL method is proposed that integrates spatial and spectral features extracted from superpixels in an AL framework. The experiments with an urban land cover classification and a wetland vegetation mapping task show that the proposed method has faster convergence and superior performance as compared to state of the art approaches. Additionally, our proposed framework has a key additional benefit in that it is able to identify and quantify feature importance-the resulting insights can be highly valuable to various remote sensing image analysis tasks.