Food Quality Inspection and Sorting Using Machine Vision, Machine Learning and Robotics

Abstract

Abstract The integration of Industry 4.0 technologies, such as machine vision, machine learning, and robotics, has transformed the food industry by enabling more efficient, accurate, and productive food quality inspection. This scope of this work is to implement these technologies in the inspection of chocolate chip cookies in a lab based setting. The PC-based system includes a conveyor belt, webcam, low-cost robotic arm, grayscale sensors, and MATLAB and Arduino microcontroller for communication between the vision system and the robot. Machine vision captures images of cookies and processes these images for feature extraction, while machine learning algorithms classify cookies based on their visual features and identify defects. The use of Artificial Neural Networks for training and testing results in an overall accuracy of 95% and 90%, respectively. The sorting of cookies based on the machine learning classification is carried out using a robotic arm. The robotic arm receives signals from the ML algorithm to remove defective cookies from the conveyor into a rejected cookies bin. The closed-loop system effectively (98%) inspects food quality, reducing defective food from reaching consumers. Machine vision and machine learning techniques offer a promising approach to improving quality control in the food industry.