Non-intrusive image processing Thompson orange grading methods

TitleNon-intrusive image processing Thompson orange grading methods
Publication TypeConference Paper
Year of Publication2017
AuthorsSabzi, S., Y. Abbaspour-Gilandeh, and J. I. Arribas
Conference Name2017 56th FITCE Congress
PublisherIEEE
Abstract

A key issue in fruit export is classification and sorting for marketing. In this work image processing techniques are used to grad Thompson orange fruit. For this purpose, fourteen parameters were extracted, comprising area, eccentricity, perimeter, length/area, blue value, green value, red value, width, contrast, texture, width/area, width/length, roughness, and length. Adaptive neuro fuzzy inference system (ANFIS), linear, and nonlinear regression methods were used. Based on results, mean square error (MSB), sum squared error (SSE) and coefficient of determination (R 2 ) were 3.47e-08, 3.47e-07, 0.988 (ANFIS), 51.33, 4927.59, 0.866 (linear reg.) and 64.85, 6092.5, 0.832 (non-linear reg.), respectively. ANFIS model was shown as the best fit model based on previously listed performance evaluation criteria.

URLhttps://ieeexplore.ieee.org/abstract/document/8093004
DOI10.1109/FITCE.2017.8093004