Computer vision based analysis of automatic indentification and sorting of recyclable beverage cans

Abstract

COMPUTER VISION BASED ANALYSIS OF AUTOMATIC IDENTIFICATION AND SORTING OF RECYCLABLE BEVERAGE ..???. CANS Irsyadi Yani, Hassan Basri, Edgar Scavino, M. A. Hannan, Noor Ezlin binti Ahmad Basri & Djuraidah A.Wahab Faculty of Engineering and Built Environment, University Kebangsaan Malaysia, 43600 UKM Bangi, Selangor DE, Malaysia ABSTRACT In recent years, identification and sorting of recyclabel beverage cans was done manually, which is still sorted by hand. Automatic identification and sorting by computer vision is the separation of object based on the pattern identification methods. For the identification processes a contact free sensor is used. After inspection and data processing a pneumatic mechanism ejects the identified particles from the belt conveyor. Separation systems that use web cameras are widely applied in different sorting applications, such as aluminum beverage cans, non-aluminum beverage cans, ans anothe metal sorting. In view of that, a study was proposed to determine the viability of using of the computer vision for the automatic identification and sorting of beverage cans and another meta1.The system achieves its ability by means of learning by examples. The training is performed using active leaming method to minimize the amount of data used in training. Experimental result shows that accuracy of the system strongly depends on the quality and quantity of the data used in tarining and lighting problems. The proposed technique shows ability to perform the classification of beverage cans with more than 90% accuracy was obtained from this research. Keywords: Computer Vision, objectdetection, Beverage cans, Identification, sorting