PAH Detection in Road Surface Residue Using Various Hyperspectral Imaging Sensors

Road destruction sites still yield tar-contaminated rubble, a material that was used until the mid 1980s in many European countries and has been shown to be carcinogenic and mutagenic. Usually, this tar-contaminated road rubble is mixed with bitumen and minerals, which are also present in road construction and are valuable for recycling. Therefore, a challenge lies in distinguishing tar from bitumen and being able to sort them for further treatment and recycling. In this work, we investigate whether it is possible to distinguish tar- from bitumen-based road surface material with hyperspectral imaging in the short and mid-wave infrared spectrum. The findings reveal that SWIR imaging is less effective, while MWIR imaging provides more promising results by distinguishing between different sources of samples based on their chemical signatures. However, generalizability across various sources poses a significant challenge due to the variability in sample composition, which is heavily influenced by the presence of different additives. Notably, the study demonstrates the efficacy of MWIR hyperspectral imaging in detecting polycyclic aromatic hydrocarbons (PAHs) in controlled sand samples, suggesting potential for identifying PAH contamination
in road debris. The results highlight several challenges and emphasize the necessity for a more diverse dataset to enhance model robustness and accuracy.