Evaluation of Minimum Preparation Sampling Strategies for Sugarcane Quality Prediction by vis-NIR Spectroscopy

Area of Science:

• Agricultural Engineering
• Spectroscopy
• Analytical Chemistry

Background:

• Proximal sensing for sugarcane quality assessment is crucial for efficient harvesting.
• Sample preparation significantly impacts the accuracy of spectral analysis.
• Visible and near-infrared (vis-NIR) spectroscopy shows potential for non-destructive quality evaluation.

Purpose of the Study:

• To identify the optimal sugarcane sample type for vis-NIR spectral analysis.
• To evaluate different sample preparation methods for predicting key quality parameters.
• To enhance the development of on-board sugarcane quality monitoring systems.

Main Methods:

• Collected sugarcane billet samples for spectral data acquisition.
• Evaluated four sample preparation modes: outer-surface (SS), cross-sectional scanning (CSS), defibrated cane (DF), and raw juice (RJ).
• Utilized Partial Least Square Regression (PLSR) to build predictive models for Brix, Pol, fibre, and TRS using vis-NIR spectra.

Main Results:

• No significant accuracy differences were found between SS/CSS and DF/RJ samples for most quality parameters.
• Defibrated cane (DF) samples demonstrated superior predictive performance for major sugarcane quality parameters.
• DF samples required minimal sample preparation, simplifying the analysis process.

Conclusions:

• Defibrated cane (DF) is the recommended sample type for accurate sugarcane quality prediction via vis-NIR spectroscopy.
• Minimal sample preparation, as with DF, is key for practical on-board monitoring.
• This study provides improved sampling strategies for efficient sugarcane quality assessment.