Feasibility of VHR Remote Sensing for Estimation of Leaf Chlorophyll Content in Armenian Vineyards

Le résumé fourni par la source

The study investigates the estimation of leaf chlorophyll content using UAV-based multispectral imagery, field spectroscopy, and in situ measurements in Trinity Canyon Vineyards, Armenia. This research aimed to enhance the accuracy of leaf chlorophyll estimation by applying and comparing advanced machine learning algorithms across different phenological stages, to assess the effectiveness of different data input methodologies for chlorophyll estimation and, therefore, determine the optimal model and input configuration for robust, spatially consistent chlorophyll quantification. Several vegetation indices were calculated and analyzed via Pearson correlation. Advanced Machine Learning models were applied to estimate leaf chlorophyll content. Results showed that Random Forest model achieved the highest accuracy during July for UAV-based 3x3 pixel box aggregation data. The findings confirm the potential of UAV-based multispectral imagery for accurate estimation of leaf chlorophyll content in vineyards.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.