Mapping of the Canopy Openings in Mixed Beech–Fir Forest at Sentinel-2 Subpixel Level Using UAV and Machine Learning Approach

Abstract

The presented study demonstrates a bi-sensor approach suitable for rapid and precise up-to-date mapping of forest canopy gaps for the larger spatial extent. The approach makes use of Unmanned Aerial Vehicle (UAV) red, green and blue (RGB) images on smaller areas for highly precise forest canopy mask creation. Sentinel-2 was used as a scaling platform for transferring information from the UAV to a wider spatial extent. Various approaches to an improvement in the predictive performance were examined: (I) the highest R2 of the single satellite index was 0.57, (II) the highest R2 using multiple features obtained from the single-date, S-2 image was 0.624, and (III) the highest R2 on the multitemporal set of S-2 images was 0.697. Satellite indices such as Atmospherically Resistant Vegetation Index (ARVI), Infrared Percentage Vegetation Index (IPVI), Normalized Difference Index (NDI45), Pigment-Specific Simple Ratio Index (PSSRa), Modified Chlorophyll Absorption Ratio Index (MCARI), Color Index (CI), Redness Index (RI), and Normalized Difference Turbidity Index (NDTI) were the dominant predictors in most of the Machine Learning (ML) algorithms. The more complex ML algorithms such as the Support Vector Machines (SVM), Random Forest (RF), Stochastic Gradient Boosting (GBM), Extreme Gradient Boosting (XGBoost), and Catboost that provided the best performance on the training set exhibited weaker generalization capabilities. Therefore, a simpler and more robust Elastic Net (ENET) algorithm was chosen for the final map creation.

Mapping Floods in Lowland Forest Using Sentinel-1 and Sentinel-2 Data and an Object-Based Approach

“…The paper did not aim to determine the significance of individual input elements (attributes) on the classification results. In this regard, it should be noted that the shortcoming of SVM is the lack of interpretability of the model in terms of inference with the input variables [107].…”

Modelling the Behavior of DVI and IPVI Vegetation Indices Using Multi-Temporal Remotely Sensed Data

“…DVI is an orthogonal and distance-based vegetation index while IPVI is a ratio index SRI LANKA Colombo District Kaduwela DSD, Colombo District [4], [16], [17]. A recent study has classified DVI as a slopebased vegetation index [13], [18]- [20].…”

Methods of Land Cover Classification Using Worldview-3 Satellite Images in Land Management

“…With the help of satellite images, the reflectance of various objects in the range of electromagnetic radiation that is not visible to humans became measurable [7]. The importance of satellite images that can be of different spatial, radiometric, spectral and temporal resolutions and therefore applicable for different purposes has been recognized worldwide, including in Croatia [8][9][10]. Satellite images of higher spatial resolutions, such as Worldview-3, reduce the possibility of overlapping spectral values among neighbouring pixels [11].…”