Olive-Fruit Mass and Size Estimation Using Image Analysis and Feature Modeling

Area of Science:

• Agricultural Engineering
• Computer Vision
• Image Processing

Background:

• Accurate olive-fruit characterization is crucial for industrial processing.
• Current methods for measuring olive size and mass can be labor-intensive and invasive.

Purpose of the Study:

• To develop and validate a novel image analysis methodology for estimating olive-fruit mass and size (major and minor axis length).
• To create accurate, automated, and non-invasive olive characterization models for industrial applications.

Main Methods:

• Acquisition of olive images (Picual and Arbequina varieties) under laboratory conditions.
• Development of an image segmentation algorithm utilizing mathematical morphology and statistical thresholding.
• Establishment of linear regression models correlating segmented image data with objective reference measurements for size and mass estimation.

Main Results:

• Variety-specific models demonstrated high accuracy: relative errors for Arbequina were 0.86% (major axis), 0.09% (minor axis), and 0.78% (mass); for Picual, errors were 0.03% (major axis), 0.29% (minor axis), and 2.39% (mass).
• Global models, applicable to both varieties, yielded comparable or superior performance to variety-specific models.
• The proposed method offers a low-cost, automated, and non-invasive approach to olive-fruit characterization.

Conclusions:

• The developed image analysis methodology provides a highly accurate and efficient means for estimating olive-fruit mass and dimensions.
• This approach represents a significant advancement towards automated, non-invasive olive characterization systems in industrial settings.
• The findings support the integration of this technology into olive processing chains for improved quality control and efficiency.