Quality Inspection of Bag Packaging Red Beans (Phaseolus vulgaris) Using Fuzzy Clustering Algorithm

Abstract

Food industry is currently focusing on fast and unsupervised quality inspection techniques. This paper deals with the development of new method for fast quality control of bag packaging red beans (Phaseolus vulgaris) using flatbed scanning. The proposed method combines fuzzy c means with spatial transformation (FCM_ST) to reduce FCM iteration. We used the labelled pixel, in the clustering image, for the evaluation of grain mixture in acquired image. The performance of the FCM_ST was compared to the standard FCM approach and it reveals itself very good for fast clustering and efficient detection of grain mixture. The detections accuracies of grain mixture in the bag packaging red beans (Phaseolus vulgaris) was 96% for acquired image with presence of other self-colour commercial beans type, 70% with presence of defected cotyledons beans, between 30% to 89% with presence of multi-coloured beans depending on their texture and between 10% to 15% with the presence of low discoloured red beans of same commercial type.