Date of Award

2008

Publication Type

Master Thesis

Degree Name

M.A.Sc.

Department

Electrical and Computer Engineering

First Advisor

Muscedere, Roberto

Keywords

Applied sciences

Rights

info:eu-repo/semantics/openAccess

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Abstract

Pharmaphil Inc. manufactures two-part gelatin capsules for the pharmaceutical industry. Their current methods of quality control of their product is by performing manual inspection of every carton of capsules prior to shipment. In today's modern manufacturing world, more efficient and cost-effective means of quality control exist. It is Pharmaphil's desire to develop a custom machine vision system to replace manual inspection with a potential opportunity in the capsule manufacturing quality control market. In collaboration with the Electrical and Computer Engineering Department at the University of Windsor, a novel system was developed to achieve this goal. The objective was to develop a system capable of inspecting 1000 capsules per minute with the ability to detect holes, cracks, dents, bubble, double caps and incorrect colour or size. Using an antiquated machine vision system for capsule inspection from the mid-nineties as a base, a modern inspection system was developed that performed faster and more thorough inspections. As a measure to minimize the overall system cost as well as to increase flexibility, a full custom design was undertaken. The resulting system follows a traditional machine vision system whereby the main components include an image acquisition component, a processing unit and machine control. The designed system uses custom USB2.0 cameras to acquire images, a standard desktop PC to process image data and a custom machine control board to perform machine control and timing. The system operates with four identical quadrants operating in parallel to increase throughput. The final system developed provided a proof-of-concept for the approach taken. The machine control and image acquisition component of the system yielded a maximum throughput of 1200 capsules per minute. After incorporating image inspection, the final result was a system that was capable of inspecting capsules at a rate of about 800 capsules per minute with high accuracy. With optimizations, the system throughput can be further improved. The findings throughout the development of the prototype system provide an excellent basis from which the first generation commercial unit can be designed.

Share

COinS