Algorithmic development in multi-view spectral x-ray imaging dedicated to waste sorting and valorization

X-ray imaging offers numerous advantages for material identification, as it is not sensitive to surface contamination, object color, and is not easily affected by their shape. Spectral X-ray imaging allows for material identification due to the strong correlation between transmission measurements and the effective atomic number of materials, potentially using multiple energy thresholds. Multi-view X-ray imaging enhances material classification by integrating information obtained from different angles. The objective of this post-doctoral research is to develop a data processing algorithm for spectral and multi-view X-ray imaging. To achieve this, it will be necessary to design the control method by leveraging simulation tools developed within CEA List, such as the CIVA simulation and analysis platform. Additionally, an algorithm for material discrimination and classification based on object shape, dedicated to a multi-view system, will need to be developed. Subsequently, the expected system performance will be evaluated through simulation, and these performances will be validated through experimental trials.