Carlos Clynes and Ita Clynes
Petroservicios, Olmo 206 Tampico, Mexico; clynes@petroservicios.com.mx
Sugarcane companies face challenges in optimizing their harvest timing, as a critical date exists where harvesting achieves maximum efficiency. Missing this date by even a few days can lead to substantial revenue losses. Traditional wet chemistry analyses of limited stalk samples lack representativeness, leading to suboptimal decisions. Industry 4.0T promotes the integration of advanced technologies such as NIR, internet connectivity (IoT) and the processing of data in cloud base type solutions to transform the way sugarcane industry operates, evaluates and produce sugar. This study evaluates the use of a portable near-infrared (NIR) spectrometer (MicroNIR Onsite-W) for in-field sugarcane analysis. A total of 1,227 stalks of various ages, varieties, and geographic origins (Honduras, Nicaragua, Guatemala) were analyzed to develop calibration models for Brix, Pol, and moisture. Chemometric models using Partial Least Squares (PLS) regression were built using laboratory reference values, achieving R² > 87% for Brix and Pol, and slightly lower for moisture. Prediction errors were 0.9%, 1.1%, and 1.2% for Brix, Pol, and moisture, respectively. This methodology enables sugarcane companies to analyze thousands of stalks directly in the field, providing a more representative understanding of plantation variability. The results save time and resources, allowing for informed harvest decisions and supporting the adoption of precision agriculture practices.