Real-Time Quality Assurance of Fruits and Vegetables with Artificial Intelligence

Abstract

Vegetables and fruits make up a major part of the human diet and finding a good grade of its quality is now a major issue in the market. To find a grade of the vegetable or fruit will be based on some parameters like size, shape, appearance, etc. The appearance is now a deciding factor for the market and affects the consumer's choice. So, we have designed an application that will classify fruits and grade them according to their quality with appearance as a parameter. This paper will describe the process which is involved in the application. This proposed system will use image processing to classify and grade the quality of fruits and vegetables by extracting features such as color, shape, and HOG (Histogram of Gradient) to classify the given fruit or vegetable. Image pre-processing techniques like data-augmentation and normalization along with Principle- Component Analysis (PCA), and also Deep learning (CNN) are used for getting good accuracy and for dimensional reduction. To speed up the identification and increase the usability, compared to current manual systems like a person checking every fruit and vegetable to grade which takes more time and energy, or by using embedded systems (sensors), we opted for a high-performance android application for quicker deployment.