A Master of Science Thesis in Mechatronics Engineering Submitted by Yahia Tachwali, "Automatic Plastic Bottle Classification System for Recycling," December 2005.Advisor for the Thesis is Dr. Abdul - Rahman Al Ali. Available are Both Soft and Hard Copies of the Thesis.
This thesis presents the design, development, implementation and testing of automatic plastic bottles sorting and classification system. The sorting is based on the bottle material chemical composition as well as on the bottle color. Sorted plastic bottles have many industrial applications. The system has two architectures; one of them is the hardware architecture which consists of a near infrared detection system and a vision system based on a charged coupled device (CCD) camera. The other one is the software architecture which is composed of two classification modules. The first one is based on the near infrared sensor and developed to sort the bottles into three classes based on the bottles' chemical composition, namely Polyethylene Terephthalate containers such as soft drink bottles, soft high-density Polyethylene containers such as milk and juice bottles, and rigid high-density Polyethylene containers such as motor oil and bleach containers. The second stage of classification is based on the CCD camera and developed to separate each near infrared system output class based on bottle color such as clear, green, yellow and gray. For each stage of classification, appropriate features are extracted to distinguish between the bottles' chemical composition or color. Consequently, various types of classifiers (namely tree classifier and quadratic discriminant function based classifiers) are developed to classify bottles based on composition and color. The final outcome of this research is a plastic recycling station that can sort and classify plastic bottles based on their chemical composition and color.