ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Publications Copernicus
Articles | Volume V-3-2020
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., V-3-2020, 193–199, 2020
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., V-3-2020, 193–199, 2020

  03 Aug 2020

03 Aug 2020


C. Sothe1, L. E. C. la Rosa2, C. M. de Almeida3, A. Gonsamo1, M. B. Schimalski4, J. D. B. Castro2, R. Q. Feitosa2, M. Dalponte5, C. L. Lima6, V. Liesenberg4, G. T. Miyoshi7, and A. M. G. Tommaselli7 C. Sothe et al.
  • 1School of Geography and Earth Sciences, M cMaster University, Canada
  • 2Department of Electrical Engineering, Pontifical Catholic University of Rio de Janeiro, Brazil
  • 3Division of Remote Sensing, National Institute for Space Research, Brazil
  • 4Department of Forest Engineering, Santa Catarina State University, Brazil
  • 5Department of Sustainable Agro Agro-Ecosystems and Bioresources, Fondazione Edmund Mach, Italy
  • 6Department of Geography, Santa Catarina State University, Brazil
  • 7Department of Cartography, São Paulo State University, Brazil

Keywords: Tropical diversity, Unmanned aerial vehicles, Deep learning, Convolutional neural network, Support vector machine, Random forest, Data augmentation, Feature extraction

Abstract. The classification of tree species can significantly benefit from high spatial and spectral information acquired by unmanned aerial vehicles (UAVs) associated with advanced feature extraction and classification methods. Different from the traditional feature extraction methods, that highly depend on user’s knowledge, the convolutional neural network (CNN)-based method can automatically learn and extract the spatial-related features layer by layer. However, in order to capture significant features of the data, the CNN classifier requires a large number of training samples, which are hardly available when dealing with tree species in tropical forests. This study investigated the following topics concerning the classification of 14 tree species in a subtropical forest area of Southern Brazil: i) the performance of the CNN method associated with a previous step to increase and balance the sample set (data augmentation) for tree species classification as compared to the conventional machine learning methods support vector machine (SVM) and random forest (RF) using the original training data; ii) the performance of the SVM and RF classifiers when associated with a data augmentation step and spatial features extracted from a CNN. Results showed that the CNN classifier outperformed the conventional SVM and RF classifiers, reaching an overall accuracy (OA) of 84.37% and Kappa of 0.82. The SVM and RF had a poor accuracy with the original spectral bands (OA 62.67% and 59.24%) but presented an increase between 14% and 21% in OA when associated with a data augmentation and spatial features extracted from a CNN.