Image processing is already established in many areas of industrial metrology. In addition to non-contact measurement, image processing is opening up more and more fields of application. In addition to the actual measurement, the goals are interaction, interpretation and ultimately understanding of the environment. This is called "machine vision", which is widely used in robotics, autonomous driving and user-friendly processing of image information, for example in medical technology. Finding features in images, whether using classical mathematical methods or methods from the field of machine learning, allows segmenting, describing and retrieving images as well as using the camera as a sensor, whether for localization or the precise determination of distances and speeds.
The lecture Digital Image Processing and the Image Processing Laboratory introduces students to the challenges of image processing and machine vision and presents solution methods for a variety of different problems.
In the laboratory, students have the opportunity to write their own image processing programs. Matlab is used, but alternatively frameworks in Python, Java or C are also provided. By means of exemplary tasks and programming exercises, the more complex topics of image processing are introduced.
The laboratory's topics are:
- Optics and depth of field
- Colour reproduction
- Linear filters in the local and frequency range
- Morphological operators and edge extraction
- Image mosaic creation using feature detectors and descriptors
The hardware equipment offers various cameras with different interfaces and lenses as well as telecentric lenses for precise industrial measuring tasks. Laboratory computers with Windows and Linux operating systems allow flexible testing of developed algorithms.
The laboratory also offers a wide range of tasks for bachelor and master theses (within the university and in cooperation with industrial companies).