Robust person recognition in 3D measurement data
For the safe detection of persons in automated logistics processes, the hazardous areas can be recorded with 3D multilayer scanners and the three-dimensional measurement data can be evaluated accordingly. Among other approaches, Principal Component Analysis (PCA) will be implemented and characterized for this purpose.
The aim of the work is to evaluate the probability with which persons can be automatically detected in a defined area.
Contents of the thesis:
- Classification of measurement data as training data
- Implementation of principal component analysis for application with 3D measurement data
- Selection of test data for statistical investigation of person recognition
- Fusion of the analysis results with the results of other methods using ensemble learning and Bayes statistics
- Programming skills (Python or MATLAB)
- Pleasure in working with data
- Independent & responsible work