Potential of SAR-derived features for detecting structural variations in coffee plots

Abstract. Brazil is the world's top producer and exporter of coffee, making it a key player in the global supply chain. This study investigates the potential of C-band Synthetic Aperture Radar (C-SAR) imagery from Sentinel-1 to detect structural differences in coffee plantations with varying cultivars and planting ages. A two-year time series of monthly radar images from Sentinel-1B and optical images from Multispectral Instrument (MSI) abord Sentinel-2 were used to compare radar-derived polarimetric features with the optical Normalized Difference Vegetation Index (NDVI). The research was conducted at Fazenda Juliana, a commercial coffee farm in Minas Gerais, which provided detailed plot boundaries and planting records. SAR data were processed using the Sentinel Application Platform (SNAP) to generate backscatter coefficients (𝜎⁰) in VH and VV polarizations. Four polarimetric indices, Cross Ratio (CR), Normalization Ratio (NL), Radar Gap Index (RGI), and Radar Vegetation Index (RVI) were calculated. These SAR features and NDVI values were segmented by plot boundaries, and average values were extracted for each plot. The temporal analysis revealed a slight cyclical trend in both radar backscatter and NDVI across the plots. A consistent positive correlation was observed between 𝜎_𝑉𝑉 ⁰ and NDVI, especially in plots with atypical vegetation responses. On specific date, NDVI fluctuations showed a strong correlation (r > 0.9) with 𝜎⁰ and NL, indicating their potential to capture structural variation in coffee crops. These results also demonstrate that radar-based features, particularly when optical data is limited, offer a reliable method for monitoring crop variability and plantation structure.