An Innovative Approach For Fruit Ripeness Classification

An automatic vision-based system for sorting and analyzing strawberry fruit is discussed in this paper. Manual classification leads to discrete opinion and is more time-consuming. Hence, finding a cost-effective and highly accurate method for strawberry classification yet remains a problem. In this paper, an automated system is proposed to predict the ripeness level of strawberry fruit, using a convolutional neural network. For the success of classification, appropriate features must be extracted. Surface color, size and shape are the necessary features for classification. The surface color of strawberry fruit determines its ripeness level. CNN is used to extract color, size and shape features from strawberry surfaces. The system shows 91.6% accuracy.