Auto Focus
Manufacturing engineering requires the ability to plan the practices of manufacturing; to research and develop tools, processes, machines and equipment; and to integrate the facilities and systems for producing quality products with the optimum expenditure of capital.[1] One of the most effective methods of developing quality technology, in a cheap and effective manner, is Roll-to-Roll processing. Roll-to-Roll or R2R processing is a fabrication method used in manufacturing that embeds, coats, prints, or laminates of varying applications onto a flexible rolled substrate material as that material is fed continuously from one roller on to another.[2]
R2R technology is used across a variety of industries, including electronics, biosciences, and materials production. Because R2R allows for large scale production of material at extremely low costs, it is the first choice of many engineering manufacturers. The only limiting factor in the already highly efficient process is production quality control. With the speeds that some manufacturers would like to develop their technology, missing defects on materials transferred from one roll to another could be catastrophic to the operation.
Our work aims to utilize imaging as the method of quality control in the R2R process. Due to the nature of the R2R process and the wide variety of materials being produced, a camera that can properly identify a defect in the transfer process with a precisely focused image is a necessity. Our goal was to construct a device that uses a stepper motor to autofocus an image taken from a Basler camera controlled with MATLAB. Both the camera and stepper motor were fit with 3d printed gears and mounted on an aluminum framing for evaluating this method of quality control.
The stepper motor driver is mounted on an Arduino and integrated with MATLAB to drive the motor. The device took advantage of a MATLAB focus measure algorithm in order to quantitatively evaluate the degree of focus of the image. This algorithm was integrated into the control loop to maximize the focus score to further the device’s performance. Additionally, a cropping feature was introduced to provide the operator with the ability to focus on a specific region of interest.
The integration of the camera, stepper motor, and MATLAB showed promise for use in an industrial setting. The device would be best suited for detecting defects in production by making use of machine learning algorithms that our research team is exploring. For the device to be used properly in engineering manufacturing it needs to share the functionality of the initial device, while still being practical and serviceable by whoever may be using it. For this reason, a mechanical redesign was needed.
The new device should be designed for ease of use in manufacturing operations while providing similar, if not better, functionality than the initial design. Because engineering manufacturing attempts to develop the best product at the optimum expenditure of capital, the new device should be not of more cost to the user than before. For this reason, a fixed frame camera device that is able to be assembled through the 3d printing of its constituent components is ideal. The new design takes advantage of fitted parts to deliver high-quality image focusing while reducing mechanical distortion. The low cost of small 3d printed parts, easy assembly, and future optimal performance simplifies R2R production quality and could make R2R the leading manufacturing method globally.
1. Matisoff, Bernard S. Handbook of Electronics Manufacturing Engineering. Springer Science+Business Media B.V., 1986.
2. Posted by Bryon Williams & filed under Article Library, Montalvo News. “Roll-to-Roll Processing: What It Is & How It Works - Montalvo.” Montalvo Corporation, Montalvo, 2 Jan. 2020, www.montalvo.com/article-library/roll-to-roll-processing-basics/.
Auto Focus
Category
Poster Presentation
Description
Session: 16-01-01 National Science Foundation Posters - On Demand
ASME Paper Number: IMECE2020-24871
Session Start Time: ,
Presenting Author: Peter DiMeo
Presenting Author Bio: Hello, my name is Peter DiMeo and I am a driven, dependable, mechanical engineering junior with a strong passion for design. Through my coursework, internship experience, personal projects, and extracurriculars I have demonstrated leadership, communication skills, and a strong work ethic. Being a former Divsion 1 Student-Athlete I have transferred my drive toward research and academia. Using the skills I have learned from the research I did this summer, I am writing a journal paper to file for an international patent, and launch a start up!
Authors: Peter Dimeo University of Massachusetts Amherst
Dan Emerson University of Massachusetts Amherst
Xian Du University of Massachusetts Amherst
Samuel Homand University of Massachusetts Amherst