Plastic Content Detection Based on ResNet and Hyperspectrum

Plastic is a widely-utilized material that generates substantial waste and pollution due to its improper disposal and recycling. Therefore, it is crucial to identify and manage plastic waste. However, existing studies primarily focuses on identifying plastic waste categories, but lacks investigation into detecting plastic particle content. Hyperspectral imaging is an emerging technology that obtains spatial and critical spectral information from images, which is highly valuable for plastic detection and identification. Nonetheless, obtaining hyperspectral images is challenging, and labeling plastic particles is problematic. This study presents a methodology for data generation and a ResNet-based framework to identify the plastic content in mixed plastic hyperspectral imagery. The feasibility of this approach is demonstrated through experimentation.