A Machine Vision Rapid Method to Determine the Ripeness Degree of Olive Lots

Area of Science:

• Agricultural Science
• Computer Vision
• Food Science

Background:

• Olive maturation significantly impacts the organoleptic quality of both oil and table olives.
• The traditional Jaén index, while indicative, is slow and subjective, lacking objectivity in determining olive ripening stages.
• Objective and efficient methods are needed to assess olive maturity for optimal harvest timing.

Purpose of the Study:

• To develop and validate a real-time, repeatable, and objective machine vision system for determining olive ripeness.
• To utilize RGB image analysis and a k-nearest neighbors algorithm for classifying olive maturation stages.
• To provide a practical tool for field-based olive ripeness assessment.

Main Methods:

• Employed RGB image analysis with a k-nearest neighbors (k-NN) classification algorithm to assess olive maturation.
• Implemented an advanced 3D algorithm for automatic colorimetric calibration to address varying lighting conditions.
• Validated the machine vision method by comparing its performance against visual evaluations by two human operators.

Main Results:

• The machine vision method achieved an accuracy of 60% in determining olive ripeness.
• The system demonstrated objectivity and repeatability, overcoming limitations of the traditional Jaén index.
• The developed algorithm provides a basis for automated, real-time assessment of olive maturity.

Conclusions:

• The proposed machine vision method offers a viable, objective, and real-time alternative for assessing olive ripeness.
• The system's potential for mobile app integration facilitates field-based, georeferenced data analysis for olive cultivation.
• Further development could enhance accuracy and expand applications in olive quality control.