A deep learning approach for automatic mapping of poplar plantations using Sentinel-2 imagery. Issue 8 (17th November 2021)

• Record Type: Journal Article
• Authors: D'Amico, G.; Francini, S.; Giannetti, F.; Vangi, E.; Travaglini, D.; Chianucci, F.; Mattioli, W.; Grotti, M.; Puletti, N.; Corona, P.; Chirici, G.
• Is Part Of: GIScience & remote sensing. Volume 58:Issue 8(2021)
• Page Start: 1352
• Page End: 1368
• Publication Date: 2021-11-17
• Subjects: Big data -- multitemporal classification -- fully connected neural networks -- forest tree crops -- tree species mapping -- deep learning; Geodesy -- Periodicals; Cartography -- Periodicals; Aerial photogrammetry -- Periodicals; Remote sensing -- Periodicals; 526.05
• DOI: 10.1080/15481603.2021.1988427
• Languages: English
• ISSNs: 1548-1603

Abstract

ABSTRACT: Poplars are one of the most widespread fast-growing tree species used for forest plantations. Owing to their distinct features (fast growth and short rotation) and the dependency on the timber price market, poplar plantations are characterized by large inter-annual fluctuations in their extent and distribution. Therefore, monitoring poplar plantations requires a frequent update of information – not feasible by National Forest Inventories due to their periodicity – achievable by remote sensing systems applications. In particular, the new Sentinel-2 mission, with a revisiting period of 5 days, represents a potentially efficient tool for meeting this need. In this paper, we present a deep learning approach for mapping poplar plantations using Sentinel-2 time series. A reference dataset of poplar plantations was available for a large study area of more than 46, 000 km 2 in Northern Italy and served as training and testing data. Two classification methods were compared: (1) a fully connected neural network (also called multilayer perceptron), and (2) a traditional logistic regression. The performance of the two approaches was estimated through bootstrapping procedure with a confidence interval of 99%. Results indicated for deep learning an omission error rate of 2.77%±2.76%, showing improvements compared to logistic regression, omission error rate = 8.91%±4.79%.