Implementation of machine-learning classification in remote sensing: an applied review

• Published in: International journal of remote sensing Vol. 39; no. 9; pp. 2784 - 2817
• Main Authors: Maxwell, Aaron E., Warner, Timothy A., Fang, Fang
• Format: Journal Article
• Language: English
• Published: London Taylor & Francis 03.05.2018; Taylor & Francis Ltd
• Subjects: Artificial intelligence; Artificial neural networks; Capacity; Classification; Data; data collection; Data requirements; decision support systems; Decision trees; Image classification; Imagery; Land-cover classification; land-cover mapping; Machine learning; Mathematical models; Neural networks; Remote sensing; Support vector machines; Training
• ISSN: 0143-1161; 1366-5901; 1366-5901
• DOI: 10.1080/01431161.2018.1433343

Machine learning offers the potential for effective and efficient classification of remotely sensed imagery. The strengths of machine learning include the capacity to handle data of high dimensionality and to map classes with very complex characteristics. Nevertheless, implementing a machine-learning classification is not straightforward, and the literature provides conflicting advice regarding many key issues. This article therefore provides an overview of machine learning from an applied perspective. We focus on the relatively mature methods of support vector machines, single decision trees (DTs), Random Forests, boosted DTs, artificial neural networks, and k-nearest neighbours (k-NN). Issues considered include the choice of algorithm, training data requirements, user-defined parameter selection and optimization, feature space impacts and reduction, and computational costs. We illustrate these issues through applying machine-learning classification to two publically available remotely sensed data sets.