Neural networks vision system for an automated tuna quality sorter

• Main Author: Bautista, Jose Jaime M., Jr.
• Format: text
• Language: English
• Published: Animo Repository 2001
• Subjects: Neural networks (Computer science); Automation; Image processing > Digital techniques; Sorting devices; Assembly-line methods; Manufacturing processes; Engineering Education
• Online Access: https://animorepository.dlsu.edu.ph/etd_masteral/2608
• Institution: De La Salle University

This paper aims to develop the applicability of using Vision System and Neural Networks using a PC-based software to control an off-line multi-input, multi-input automated tuna quality sorting system. Initially, tuna is loaded on a belt conveyor running at a constant speed. Each time a tuna passes a first set of photo-sensors, a video camera installed overhead is triggered to capture its image. The captured image of the tuna is transformed into a data using image-processing algorithms, processed and analyzed by an IBM-PC that determines the quality of the tuna. As the belt conveyor continuously moves toward the end of the line, the tuna again passes another set of sensors that activate the flipper arm to move on the desired position and allow the tuna to slide to its corresponding bin.