Applying Convolutional Neural Networks to Land Cover and Groundwater Potential Domain Classification

Abstract

In the field of groundwater engineering, Convolutional Neural Networks (CNNs) are playing an important role in evaluating spatial groundwater potential domains and land use/land cover changes based on remote sensing (RS) technology. CNNs can offer great potential for extracting complex spatial features with several advanced generalizations. However, geometric distortions and blurry object boundaries and massive data preparation disruptions can become major constraints and affect the spatial potential of CNNs for land cover classification.