Monitoring quality of intact olives with near-infrared spectroscopy

Significance of the topic

Near-infrared spectroscopy (NIRS) offers a rapid, non-destructive and reagent-free approach for assessing key quality attributes of intact olives prior to oil extraction. Determining oil content, moisture and maturity directly on whole fruit enables better harvest scheduling, process planning at the mill and faster quality control compared with traditional wet-chemistry methods that are time-consuming and solvent-intensive. Implementing robust NIR models in production environments can improve yield estimates and economic returns while reducing laboratory workload.

Objectives and study overview

This application note (AN-NIR-153) evaluated the feasibility and performance of NIRS to predict oil content, moisture content and a maturity index on intact olives. The study used a large sample set representing two cultivars (Picual and Arbequina) across varying ripeness stages to build and validate quantitative prediction models suitable for routine analysis in oil mills or QC laboratories.

Methodology

Key experimental design points:

• Sample set: 800 intact olive samples comprising two cultivars (Picual and Arbequina) with a range of oil and moisture levels and maturity stages.
• Spectral acquisition: Diffuse reflectance NIR spectra collected from 1000 to 2250 nm using rotating and multipoint measurements to address sample inhomogeneity.
• Measurement protocol: Multiple positions per sample and sample rotation to average heterogeneity; no sample preparation (no grinding).
• Calibration/validation: 75% of samples used for calibration and 25% for validation. A leave-one-out cross-validation procedure was applied during model development.
• Reference methods: Oil content by Soxhlet extraction; moisture by loss-on-drying; maturity index according to International Olive Council (IOC) standards.

Used instrumentation

Equipment and software employed in the study:

• OMNIS NIR Analyzer Solid (diffuse reflectance, 1000–2250 nm) — Article No. 2.1071.0010
• Large holder OMNIS NIR, 100 mm — Article No. 6.07402.100
• Large cup OMNIS NIR, 100 mm — Article No. 6.07402.110
• OMNIS Stand-Alone software license — Article No. 6.06003.010
• Software license Quant Development for model building — Article No. 6.06008.002

Main results and discussion

NIR predictive models were developed for oil content, moisture content and maturity index. Spectra from intact olives produced models with the following figures of merit (FOM):

• Oil content (reference: Soxhlet): R2 = 0.811; SEC = 1.39%; SECV = 1.43%; SEP = 1.44%.
• Moisture content (reference: loss on drying): R2 = 0.868; SEC = 1.70%; SECV = 1.75%; SEP = 1.81%.
• Maturity index (reference: IOC visual/standard method): R2 = 0.706; SEC = 0.48; SECV = 0.49; SEP = 0.51.

The results indicate strong predictive performance for moisture and good performance for oil content directly on intact fruit. The maturity index model showed lower correlation, reflecting the greater challenge of predicting a visual/qualitative index from diffuse reflectance spectra and likely dependence on cultivar-specific skin and pulp characteristics. Use of rotating multipoint acquisition improved representativity for non-homogeneous whole-fruit samples.

Practical considerations discussed in the study include the benefit of eliminating sample grinding and solvent extraction for routine screening, but also the need for representative calibration sets spanning varieties, ripeness stages and growing conditions to ensure model robustness. Model transfer and periodic recalibration are implied requirements for wider deployment across different mills or seasons.

Benefits and practical applications

Practical advantages demonstrated or implied by the study:

• Non-destructive, reagent-free quantification of oil and moisture in intact olives, enabling immediate feedback at harvest or at the mill intake.
• Rapid measurement times (instrument capability to measure solids in less than 10 seconds), suitable for high-throughput screening.
• Automated multi-position measurement and sample rotation to improve reproducibility with heterogeneous samples.
• Integration capability with automation systems and lab workflows for routine QC and process control.

Future trends and potential uses

Potential developments and extensions of this approach include:

• Inline or at-harvest NIR systems and portable NIR devices for field-level oil content estimation to better time harvest operations.
• Fusion with imaging or hyperspectral methods to map spatial variation within lots and enable sorting by ripeness/oil content.
• Application of advanced chemometric and machine learning algorithms to improve prediction accuracy and robustness across cultivars and seasons.
• Calibration transfer strategies and transfer learning to allow shared models across instruments and sites without full recalibration.
• Integration with mill automation to link intake quality to process parameters and optimize extraction yield and oil quality.

Conclusion

The study demonstrates that NIRS measured on intact olives using the OMNIS NIR Analyzer Solid can reliably predict oil and moisture content with performance adequate for routine screening in production and QC settings. While the maturity index model showed lower predictive power, the overall approach removes the need for destructive sample preparation and solvent-based reference analyses for many screening tasks. For broad industrial use, successful deployment requires representative calibrations, validation across seasons and cultivars, and attention to model maintenance and transfer.

Reference

Metrohm Application Note AN-NIR-153: Monitoring quality of intact olives with near-infrared spectroscopy — Determination of oil and moisture content in intact olives.
International Olive Council (IOC) Standards, Methods and Guides (maturity index reference mentioned).
Soxhlet extraction method for oil content (standard laboratory reference method).
Loss on drying method for moisture determination (standard laboratory reference method).