Using multispectral images acquired by unmanned aerial vehicle for assessing the nutritional condition of mango trees in Son La province

Abstract

Unmanned aerial vehicle (UAV) equipped with multispectral sensors is increasingly utilised for monitoring and assessing crop health, playing a crucial role in precision agriculture. This study employed the Normalised Difference Vegetation Index (NDVI) computed from images captured by Phantom 4 Multispectral UAV for rapidly assessing the nutritional status of
mango fruit trees in Yen Chau district, Son La province (before merging) across four different stages: pre-flowering, flowering, fruiting, and harvesting. The results revealed that the NDVI values of mango trees peaked during the pre-flowering stage (NDVI>0.8) but significantly decreased during the flowering and fruiting stages (NDVI from 0.6 to 0.8). Nutritional conditions (good, moderate and poor) were classified based on defined NDVI thresholds, achieving high overall accuracies, ranging from 73% to 82%. The majority of mango trees in the study area exhibited good nutritional status, accounting for approximately 90%, while the proportion of trees categorised as moderate and poor ranged from 5 to 8% and from 1 to 3%, respectively, depending on the stage. This study indicates the potential of UAV technology in monitoring and evaluating fruit trees, paving the way for integration with other digital technologies to enhance the efficiency of sustainable agriculture in Son La province.