Development of Machine Learning based Fruit Detection and Grading system

Consuming of healthy fruits is pivotal for all human beings as they are the source of energy. It is essential to consume good quality of food items. These days' buyers are assessing food visually for its quality. This manual procedure brings about additional time and it is a relentless and tiring task. Thus, there is a need for an automatic machine which identifies the imperfections, and sorts them as per quality. The proposed system captures the fruit placed on conveyor belt then the captured image is compared with the trained data set using Convolutional Neural Network Network (CNN) algorithm which extracts the features of the fruits like texture, color, and size. In the convolution layer of CNN edges from raw pixel data are detected further. These edges are used to detect shapes and then higher level features are detected by shapes. This paper presented the sorting of apple fruit based on their quality such as Type A (best) Type B (raw or average) and Type C (worst) done by using CNN which is a deep learning algorithm. The objective of the work is to improve the accuracy and efficiency by automatic sorting system which mainly helps in reducing time. The proposed model classifies the fruits with an average accuracy of 96.66%.