Evaluating the ability of convolutional neural networks for transfer learning in Pinus radiata cover predictions

A. Bravo-Diaz, Sebastián Moreno, Javier Lopatin

Abstract: Highlights•We tested transfer learning CNN capabilities for Pinus radiata regressions.•We compared transfer learning outside training data domains.•We compared UAV data with low and high spatial variability.•The transfer learning in homogeneous data depends on domain similarity.•The best transfer learning models often disagree with test-based selections.