Class-based plastic categorisation RVM
We are developing a new-generation reverse vending machine (RVM), which aids class-based plastic pre-sorting. Users can insert used plastic products (pre-cleaned), whatever the class (i.e., #1 to #7), into our RVM.
Conventional RVMs utilise only barcode scanning to identify exclusively PET bottles. But not all plastic bottles have barcodes, and not all are made of PET plastic. To achieve universal recognition of plastic classes, we use near-infrared (NIR) technology, revealing a unique spectral fingerprint under excitation of a designated NIR wavelength. At the current stage, we are establishing our own printed circuit board assembly (PCBA) model for data collection. The PCBA model receives discrete reflectance data, which is processed to construct an estimated spectral model. Then computer learning will play a role in correlating the model and the class of plastic with the help of references provided by the existing NIR spectrometers.
Team members
Mr Wu Ho-hoi (Hong Kong Baptist University)
Mr Leung Chun-hin (The Hong Kong Polytechnic University)
Mr Li Hong-lem (The Hong Kong University of Science and Technology)
* Person-in-charge
(Info based on the team's application form)
- CityU HK Tech 300 Seed Fund (2022)
- Finalist, HKSTP Hong Kong Techathon (2022)
- Innovation Award, City I&T Grand Challenge (2021)