Digital image analysis in a mobile device to estimate surface and volume of mango fruit (Mangifera indica L.)

Abstract

Objetive: To Develop a technique based on the partition method to estimate surface and volume of mango fruit (Mangifera indica L. cv. ‘Ataulfo’), through two orthogonal digital images captured and processed directly in an Android® mobile device using different discriminants for segmentation.

Design/methodology/approach: A technique to estimate surface and volume of mango fruit (cv ‘Ataulfo’) was developed. Images were captured and processed directly in an Android mobile device. Three discriminant for segmentation were evaluated (red channel, green channel and HSI color intensity). Technique was applied to 30 mango fruit. Precision and accuracy were determined.

Results: Surface and volume estimated using the red channel showed a coefficient of variation less that 2.0%. While estimation error were less that 3.0% for surface and less that 5.0% for volume.

Study limitations/implications: Technique is limited to fruit with elliptical transverse section and smooth surface.

Discoveries/conclusions: The proposed technique is a non destructive alternative to estimate surface and volume of mango fruit with reasonable error margin. It can be directly applied in an Android mobile device using as discriminant to segmentation only the red channel.