Biomass estimation of abandoned orange trees using UAV-SFM 3D points

Abstract

In smallholder areas, the abandonment of orchards is a recent phenomenon with socioeconomic and environmental consequences. Biomass estimation and monitoring of these areas is essential to analyze their influence on the CO2 balance and to quantify carbon pools. In the current context of energy supply uncertainties and considering the demanding use of alternative energy sources, the quantification of fruit tree biomass in abandoned areas is a question of great interest. In this study, the above biomass of abandoned orange trees was estimated using tree parameters calculated from 3D points derived from images captured by a UAV and applying the Structure from Motion (SfM) technique. From these data, a canopy height model was calculated and used to apply a developed crown contour detection algorithm. Using this information and 3D points, the tree parameters crown area, crown diameter, crown length, maximum tree height, minimum tree height, mean and standard deviation of crown point heights were calculated for a set of 36 felled and weighted orange trees. Stepwise regression was used to estimate the above biomass values. All previously reported variables were included. The crown area parameter produced the most accurate model with R2, RMSE and RMSE % values ​​of 0.85, 10.165 kg and 19.56 %, respectively. These results demonstrate the potential of UAV-SfM-derived 3D point clouds to estimate the above-ground biomass of abandoned fruit trees, relevant information for environmental analysis and biofuel energy production.

Keywords

• Biomass
• Structure from Motion
• Abandoned crops
• CHM