Katja Kossira

The majority of plastic waste is thermally recycled, since it is usually not possible to separate it by type. To enable clear classification, I have been working on multispectral imaging since January 2023. The goal is to clearly recognize and identify plastic materials using multispectral imaging, thus ensuring recyclability.

For further information please visit:

I offer thesis in the field of Multispectral Imaging and data processing at any time. Knowledge in Python or MATLAB is desirable.

If you are interested or for further information please send me an email with a short introduction of your interests and previous experiences in the field of software development or image and data processing.

• Master Thesis
  • Conditional Optimal Filter Selection in Multispectral Imaging for Plastic Recycling (2024)

• Bachelor Thesis
  • Evaluation of State-of-the-Art Inter-Camera Spectral Calibration Techniques (2023)
  • Comparison of Interpolation Methods for the Spectral Reconstruction of Plastic Material using a Multispectral Database (2024)
  • Determination of the Spectral Variance of Plastics using a Multispectral Database (2024)
• Research Internship
  • Evaluation of the Robustness of Statistical Inter-Camera Color Calibration Techniques for Multispectral Imaging (2024)

• Bachelor Thesis
• Master Thesis
  • Development of a Neural Network for Multispectral Plastic Material Classification
• Research Internship
  • Multispectral Inter-Camera Color Correction using Weighted Average and Weighted Median Consensus Images
  • Evaluation and Comparison of different Hyperparameters for the Conditional-Filter-Band-Selection Algorithm Classifiers
  • Evaluation of the Spectral Variance of Plastics within the same Material Classes using a Multispectral Database