This project covers several aspects of the process for verifying compliance with regulations based on samples of recycled plastic material — from analyzing the applicable standards to developing new methods of sample processing using artificial intelligence techniques.

In summary, the project will include the development of:

• An online data management platform – A scalable web interface for managing experiments and checking whether results comply with different regulations, supported by a comprehensive data framework.
• A computer vision system – Acquisition and preprocessing of pellet and film images; testing and training of deep learning models; development of an image-based quality classification system; and integration of this system into the web platform.
• A quality evaluation and comparative methods module – Creation of a quality evaluation framework using PCA and other methods; experimentation with machine learning techniques; and integration and validation of this evaluation system within production.
• Technical documentation and user manuals – Development of complete technical documentation to enable future code extensions or modifications, along with user manuals and reference guides for the software.