Detection of Papaya Ripeness Using Deep Learning Approach

Abstract

Papaya is a berry fruit with nutritional as well as real worth because to its non-seasonality and short harvesting period. The fiscal year 2020 statistics shows that the volume of papaya production increased to over six million metric tons in India. The grading of papayas is done by hand by human operators, which might lead to misclassifications. The identification of the ripeness of a fruit is important in case of distributing the classified papaya packages as well as in purchasing them by customers. Many projects were proposed earlier for classifying fruits and vegetables, however they were done using machine learning algorithms while the proposed system focuses on deep learning algorithm, especially using Convolution Neural Network (CNN). Convolution Neural Network is a deep learning technique that identifies features without the need for manual absorption. The papaya dataset which is used for this system consist of 300 images, in which each class (ripe, unripe and partially ripe) has 100 images. The proposed model is expected to have a maximum accuracy.