Tropical fruits grading using magnetic resonance imaging based intelligent technique

Abstract

Abundance of variety of fruits is produced in most parts of the tropical countries. However, facilities for effective storage of these fruit are not readily available making them appear edible on the surface but in most cases the quality of their tissue and juice indicates that they are internally defective. Present procedures for assessing fruit quality are X-ray, impedance, image analysis, laser spectroscopy, sonic, ultrasonic and reflectance spectroscopy but to date no further applications are known that have taken unique advantages of Magnetic Resonance Imaging (MRI). This study, applied Magnetic resonance imaging as an effective non-destructive technique to determine the internal quality of selected tropical fruits. In this thesis, two types of tropical fruits, namely oranges and banana are considered for their internal tissue evaluation. Collected data using MRI equipment were Fourier transformed to obtain related MR images once the K-space has been fully assembled. The MR images of fruit obtained from proposed image processing methods were used as features to recognize their internal quality. The variances determined and features obtained were applied to the Artificial Neural Network (ANN) code developed in MATLAB using fruit MRI intensity features as inputs. The obtained results have shown that the Levenberg-Marquardt algorithm (LM) with Mean Square Error (MSE) and R-value of 0.0814and 0.9379 for Orange and 0.0693 and 0.9989 for Banana fruit respectively produced the best performance fitness for the assessment of internal quality of the fruits. These support the view that a combination of ANN and MRI has the capability of identifying the defect as well as being able to perform better than other non-destructive techniques.