Low Complexity Approach of Material and Relative Appearance Detection for Beverage Cup

摘要

The beverage cups influence food culture and environmental issues. EPA announced that the amount of beverage cup amounted to 1.5 billion in the 2011, and gradually pays attention to this problem. The beverage cups include the paper, PP, PS, PLA, PET etc. In this paper, we proposed an image-based recognition system to detect the beverage cup. These methodologies contained the image processing techniques and simple geometric operations to find the appearance of the cup, such as cup cover, straws, cup body. The image processing technologies contained the gray level transformation, Gaussian distribution and edge detection (Canny). Furthermore, we use the image matching approach based on Sum of Absolute Difference (SAD) to find the corresponding material. It can improve recycle efficiency, saving the labour power of beverage cup recycle, and making a contribution to the environment and recycle. The average time cost was 103ms. The accuracy rates were: straw 90%, PLA 66%, PP 95%, PS 100%, PAPER 100%.