Design and implementation of small color recognition system for beverage bottle caps

Abstract

Objective: This study focus on solving the problems that the current beverage bottle cap detection system has simple functions, large volume and low color recognition rate. Methods: A design scheme of small beverage bottle cap color recognition system based on ARM processor was proposed. By using the LineSim tool of the simulation software HyperLynx, a four-layer PCB board was designed and impedance matching simulation analysis was obtained after setting transmission line parameters. Through using the compilation software Jupyter Notebook to achieve threshold setting, target contour detection, target frame out and other functions. Results: Under bright and low light conditions, the bottle cap targets of red, green and blue were tested, and the recognition rate of the design scheme for the three colors could reached more than 92.7%. Conclusion: Compared with the Fourier descriptor, the scheme has higher recognition accuracy and precision, and the system also has the function of face recognition, which is suitable for various intelligent application scenarios.

Keywords

• machine vision
• bottle cap inspection
• impedance matching
• color recognition
• contour inspection