Hyperspectral image analysis for material recognition

Automated sorting systems are widely used in industrial applications, e.g., quality control of food and recycling of various materials. Up to know, high-speed sorting is mostly realized by optical inspection in the visible range of the electromagnetic spectrum. The classification result of challenging sorting tasks can be improved by analyzing the near-infrared reflectance spectrum. During this research project, different ways of extending the inspection systems to the near-infrared spectrum without losing speed or resolution are evaluated.

Research Objectives:

  • Hyperspectral image acquisition
  • Hyperspectral data analysis of polymers, food, and minerals
  • Band selection techniques
  • Implementation of fast classification algorithms