EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Deep Learning Approach for Pixel-Level Material Classification via Hyperspectral Imaging

Savvas Sifnaios iD, George Arvanitakis, Fotios K Konstantinidis iD, Georgios Tsimiklis, Angelos Amditis iD, Panayiotis Frangos iD

Journal of Imaging · 18.06.2026 · 12 (6) · 267

DOI: 10.3390/jimaging12060267 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Imaging, online erschienen am 18.06.2026. Verfasst von Savvas Sifnaios und 5 weiteren Autor:innen. Beteiligte Einrichtung: Institute of Communication and Computer Systems, National Technical University of Athens, 9 Iroon Polytechniou Str., GR-157 73 Athens, Greece. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 53 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: European Union.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Recent advancements in computer vision, particularly in detection, segmentation, and classification, have significantly impacted various domains. However, these advancements are still strongly tied to RGB-based systems, which are insufficient for applications in industries such as waste sorting, pharmaceuticals, and defence, where material characterization beyond shape or visible colour is necessary. Hyperspectral (HS) imaging captures spatial and spectral information for each pixel and therefore offers a promising route for material-level classification. This study evaluates the potential of combining HS imaging with deep learning for plastic material classification. The work includes: (i) the design of an experimental setup with a HS line-scan camera, conveyor, and controlled illumination; (ii) the construction of an object-disjoint dataset of HDPE, PET, PP, and PS samples with semi-automated mask generation and Raman spectroscopy-based labelling; and (iii) the development of P1CH, a lightweight pixel-wise 1D convolutional hyperspectral classifier. On object-disjoint test images, P1CH achieved 97.44% all-pixel accuracy. A boundary sensitivity analysis, reported separately because semi-automated labels are uncertain at material/background interfaces, yielded 99.94% accuracy after excluding a pre-defined two-pixel border band. Additional ablation, baseline, and robustness analyses show that the proposed pixel-wise spectral approach is effective for small fragments, visually similar plastics, and overlapping materials, while black or very dark plastics remain challenging under the present camera and illumination configuration.

Schlagwörter

Stichwörter der Autor:innen

deep learninghyperspectral imagingmaterial classificationpixel-level classificationreal-time object detection

Autor:innen und Institutionen

  • Institute of Communication and Computer Systems, National Technical University of Athens, 9 Iroon Polytechniou Str., GR-157 73 Athens, Greece.
  • School of Electrical and Computing Engineering, National Technical University of Athens, 9 Iroon Polytechniou Str., GR-157 73 Athens, Greece.
  • Technology Innovation Institute, Yas Island, Abu Dhabi P.O. Box 9639, United Arab Emirates.

ORCID-Profile: Savvas Sifnaios, Fotios K Konstantinidis, Angelos Amditis, Panayiotis Frangos

Förderung

  • European Union

Zitieren

APA

Sifnaios, S., Arvanitakis, G., Konstantinidis, F. K., Tsimiklis, G., Amditis, A., & Frangos, P. (2026). A Deep Learning Approach for Pixel-Level Material Classification via Hyperspectral Imaging. Journal of Imaging, 12(6), 267. https://doi.org/10.3390/jimaging12060267

Vancouver

Sifnaios S, Arvanitakis G, Konstantinidis FK, Tsimiklis G, Amditis A, Frangos P. A Deep Learning Approach for Pixel-Level Material Classification via Hyperspectral Imaging. Journal of Imaging. 2026;12(6):267. doi:10.3390/jimaging12060267

BibTeX
@article{Sifnaios2026,
  author = {Sifnaios, Savvas and Arvanitakis, George and Konstantinidis, Fotios K and Tsimiklis, Georgios and Amditis, Angelos and Frangos, Panayiotis},
  title = {A Deep Learning Approach for Pixel-Level Material Classification via Hyperspectral Imaging},
  journal = {Journal of Imaging},
  year = {2026},
  volume = {12},
  number = {6},
  pages = {267},
  doi = {10.3390/jimaging12060267}
}
Kostenlos starten →