Monitoring and assessing pasture condition represent a fundamental element in the sustainable management of agrosilvopastoral resources, having direct implications for productivity, livestock carrying capacity, and biodiversity conservation. In the current context of the transition towards smart agriculture, developing accessible and scalable methods for the spatio-temporal analysis of vegetation has become a priority. This paper proposes a pasture evaluation approach based exclusively on visible spectrum (RGB) imagery, demonstrating that, through the use of open-source data and deep learning algorithms, results comparable to those obtained with traditional multispectral methods can be achieved. The main objective of the study is to highlight the potential of RGB data, with a spatial resolution of 20–30 m, for determining vegetation indices and classifying pasture conditions. The analysis was conducted on grasslands located in Timiș County, Romania, using both annual and multi-annual time series derived from the Copernicus program. The methodology integrated the computation of vegetation indices specific to the visible spectrum and the application of a classification model based on convolutional neural networks (CNN). The model was trained and validated on labeled datasets, achieving an accuracy of over 95% on the test set, confirming the robustness of the proposed approach. The results revealed a significant correlation between seasonal variations of vegetation indices, precipitation patterns, and the specific phenology of the studied grasslands. The multi-annual analysis enabled the identification of degradation and regeneration trends in the grass cover, contributing to a better understanding of ecological dynamics and the impact of climatic factors on green biomass. Furthermore, the study introduces a Python-based application that integrates the entire workflow: satellite data acquisition, vegetation index computation, and CNN-based classification. Through its open, reproducible, and scalable nature, this technological solution provides a modern tool to support farmers, agronomists, and decision-makers involved in the sustainable management of pastures and the adaptation of agricultural practices to climate change.

vegetation indices; RGB imagery; CNN; grasslands monitoring; smart agriculture

agronomy

Presentation: poster