Texture and Colour Gradient Features for Grade analysis of Pomegranate and Mango Fruits using kernel-SVM Classifiers

Abstract

There is an extended demand for quality fruits and vegetables for processing into juice, wine and syrup in today's competitive market. Traditional fruit and vegetable processing and grading of good and fresh quality is time consuming and requires more skilled labour. Computer vision and Machine learning approaches are the best solutions to the above mentioned problem. The present paper implements a novel approach of grade classification of pomegranate and mango fruits with texture and colour gradient features. Texture of the fruits are modelled using structural features using local binary pattern (LBP) and statistical features using pixel run length matrix (PRLM) and GLCM, while colour gradients (CG) of the fruits are calculated using average colour gradients, variances and colour coordinates of the three primary colours red, green and blue. Kernel support vector machine (KSVM) is used to grade/classify the extracted features from the proposed and existing methods. The statistical performance results show that the proposed approach is effective in grade classification and defect identification of the fruits with varying texture and colour gradients to an acceptable degree.