Improved Fruit Detection by Image Processing using Deep Learning

In the food market today, where everyone is concerned about their health, the capability of a deep learning architecture to recognize fruits based on their quality is crucial. Fruits come in a variety of varieties. However, in order to specifically identify the fruits with the highest quality, the most effective classifier for fruit classification is found with the use of dimensional analysis during the pre-processing of images in terms of time complexity. Finally, this study discusses about how the proposed deep learning model may be applied and how it might be expanded to suit new demands. The proposed CNN-based model was loaded into the special program to test its performance on 200 different apples, and its efficiency is achieved as the measure of accuracy with 92% and with a processing time of less than 72 MS for six photos of an apple fruit. The overall results showed that there was an increasing possibility for implementing the proposed CNN-based classification model in commercial packaging lines. The proposed system starts the process by clicking the picture of the fruit. The characteristics of the fruit samples, such as their size, shape, and colour are predicted with 92.7% accuracy.