EFFECTS OF FLIGHT ALTITUDES ON THE DETERMINATION OF POMELO FRUIT TREE LOCATIONS AND HEIGHTS USING UNMANNED AERIAL VEHICLE
Keywords:
UAV; Flight altitudes; Tree locations; Tree heights; Canopy height modelAbstract
Determination of an appropriate flying altitude is a crucial task when using unmanned aerial vehicles (UAV) for tree monitoring. This study tested and evaluated the influence of different UAV flight altitudes (30 m and 50 m) on the computation of pomelo fruit tree locations and heights. The results indicated that pomelo fruit trees were detected from the UAV images captured at a flight altitude of 30 m obtained an accuracy of 92.2%, higher than data captured at a flight altitude of 50 m (87.4%). Tree heights extracted from UAV data achieved high accuracies for both flight altitudes but 30 m flight altitude produced more accurate results than 50 m flight altitude; with an mean squared error (MSE) of 0.27 versus 0.33, respectively. A total of 353 pomelo trees were identified within the study area. The average tree height was 3.3 m, with the lowest and highest trees were 0.5 and 6.2 m, respectively. Decreasing the flight altitude from 50 m to 30 m results in an increased accuracy, but not significant. Thus, UAV applications for large areas could set a height of 50 m and should still obtain reliable results. This paper will be an important reference for future studies on UAV applications for fruit tree monitoring.