Machine vision system for glass bottle inspection using LabVIEW

• Main Authors: An, Michael Vincent G., Cruz, Rommel Sebastian P., Ferrer, Geoffrey T., Sy, Leo Augustus L.
• Format: text
• Language: English
• Published: Animo Repository 2010
• Subjects: Computer vision; Process control > Automation; Machinery > Monitoring; Bottle industry
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/14682
• Institution: De La Salle University

In bottling industries of today, glass bottle inspection is done both by manual human inspectors and an automated machine. This study aimed to design and construct a glass bottle inspection system suitable for fault detection using LabVIEW, a graphical programming software. The fully automated machine vision system is able to detect chipping and cracks on the glass bottle wall and mouth. These defective glass bottles are then rejected from the conveyor. The proponents were able to construct an inspection environment that included proper lightning, placement of system components, and conveyor speed for suitable and clear image capturing and defect detection. An algorithm was designed and run in LabVIEW that synchronized the entire system operation and processed image data to evaluate a given clear glass bottle. Image processing techniques such as grayscaling, pattern matching, edge detection, and object detection were applied. An infrared proximity sensor was used for bottle detection and image acquisition devices in the form of webcams were utilized for bottle image capturing. Through experimentation, the fault detection accuracies of the system were as follows: 87% for overall bottle inspection, 75% for bottle mouth inspection, and 83% for bottle body inspection.