The Automated Sorting Methods Modernization of Municipal Solid Waste Processing System

In this paper the plastic sorting system of the waste processing line is considered. Plastic is a valuable raw material for recycling but only 10% of plastic waste is processed nowadays in Russia. Two main methods of waste recognition are applied in two separation stages: spectrometry and computer vision. The implementation of municipal solid waste sorting system using the near infrared and visible spectrum spectrometers can significantly increase the efficiency of separation process in comparison with manual sorting. But the quality of separation is still low because this plastic cannot be separated by color using this method and a lot of impurities are included. The evident conclusion is the need for two-stage of sorting process: the rough but fast sorting using spectrometry and fine sorting using optical methods. Another issues of this paper are related to optical recognition subsystem: we propose a fast multi-level recognition system, the first level is the detection of the object presence on the line and the analysis of camera focusing, the second is the classification of the plastic object using the joint solution of several neural networks. This approach helps to decrease the total operating time by estimating the necessity of neural network running and by avoiding the blurred and faulty image processing.