Exploring Multisource Remote Sensing Capacities for Spatiotemporal Analysis and Quantification of Chlorophyll of Vineyards in Armenia
Abstract
The potential of satellite and unnamed aerial vehicle (UAV) multispectral images for estimating the dynamics and the contents of leaf chlorophyll in vineyards was investigated in this research study. A series of PlanetScope images were used to implement a time series analysis for 2017–2023. Normalized Difference Vegetation Index (NDVI) and Green Normalized Difference Vegetation Index (GNDVI) were calculated to assess the spatial and temporal changes of vegetation. The Mann–Kendall trend test was used to determine if a linear monotonic trend exists in a given set of time series data. To estimate the contents of leaf chlorophyll, a UAV survey was conducted and simultaneously in situ measurements of grape leaf chlorophyll content were performed to study the relationships between UAV-derived spectral indices and in situ measured leaf chlorophyll content. Two machine learning (ML) models (random forest (RF) and partial least squares regression (PLSR)) were applied to estimate the content of the chlorophyll in grape leaves, and the results were optimized via stacked generalization. According to the Mann–Kendal trend test, “no trend” was detected for the whole study area. However, the trend varies for grape varieties for the study period, which indicates some characteristics of the development of different grape varieties. RF (R2Val = 0.63, RMSEVal = 25.80) and PLSR (R2Val = 0.64, RMSEVal = 25.56) models showed robustness for chlorophyll estimation in the studied vineyard. For the next steps, the ML will be used to show the characteristics of the spatial distribution and heterogeneity of the chlorophyll in grape leaves in Trinity Canyon Vineyards.
https://doi.org/10.1007/s40003-025-00845-8