Studies on microplastic pollution commonly use expensive spectroscopic methods such as FTIR spectroscopy or Raman spectroscopy. Alternatively, studies could also employ cheaper methods such as optical microscopes, which are prone to errors. Here, we present an inexpensive and simple multi-spectral imaging system that is based on photoluminescence spectroscopy. The main components are a RGB camera, two light sources that have a wavelength of 365 nm and 405 nm and a set of bandpass filters (with a central wavelength ranging from 380 nm to 440 nm). Our study shows that the spectra acquired with both light sources are sufficient to distinguish plastics from natural materials and identify different plastic types. Considering the simplicity of the proposed system, this could be combined with methods to extract microplastics from the environment.