Identification and Classification of Botrytis Disease in Pomegranate with Machine Learning

The Botrytis cinerea represents an economic risk for the pomegranate industry because the quality of pomegranate-derived products is mainly affected by the number of bad arils present in the fruit. Manual identification and classification of this fruit requires expertise and professional skill and is time-consuming, expensive, and subjective. Automated identification and classification of Botrytis can be an alternative to the traditional manual methods. Machine learning techniques such as K-nearest neighbor algorithms, support vector machines (SVMs), random forest, and artificial neural networks have been successfully applied in the literature for fruit classification problems. In this paper, we propose a new method to identify and classify Botrytis disease of the pomegranate through combining machine-learning techniques. The method also uses different techniques such as Gaussian filter, morphological operations, among others, to extract the image features. The results show that 96% of classification accuracy can be achieved using the proposed method.