Abstract

In recycling, highly sorted waste paper streams facilitate high quality end products and save processing chemicals and energy because various grades of waste papers are subjected to different recycling processes. Automated paper sorting systems provide more efficient and effective advantages over human inspection from worker fatigue, throughput, speed, and accuracy point of view. Consequently, many automated mechanical and optical paper sorting methods have been developed to fill the paper sorting demand during 1932 to 2009. Because of inadequate throughput and some major drawbacks of mechanical paper sorting systems, the popularity of optical paper sorting systems has increased. The implementations of the previous methods are still complex, expensive and sometimes offer limited reliability. This research attempts to develop a smart vision sensing system that is able to identify the grade of waste paper using window features. For constructing reference template database, hue and saturation of selected region in the paper object image are considered. The paper grade is identified based on the maximum occurrence of a specific reference template in the paper object image. The classification success rates with window size 3×3 for white paper, old Newsprint paper and old corrugated cardboard are 95%, 88% and 90%, respectively and then the achieved average classification success rate is 91.07%. The remarkable achievement obtained with the method is the accurate identification and dynamic sorting of all grades of papers using window features, which is the best among the prevailing techniques of optical or electronic image based systems in terms of throughput, performance in identification, adaptability with new paper grades, and cost of implementation.

Paper Details