Laboratory description
AI methods in the field of surveying and geoinformatics are applied and researched in the laboratory.
The methods and models for the application of artificial intelligence taught in lectures can be applied and tested in practice in the laboratory. For example, when analysing large and complex data sets such as satellite and aerial images, elevation models, etc..
Powerful computers with suitable graphics cards are required for the application of AI methods, especially for the training of neural networks. There are 8 GPU-based desktop computers available in the laboratory, which can be used for teaching as well as for final theses and other projects.
Automated detection of vegetation in high-resolution aerial images:
An example of the results of our segmentation model: In most cases, the model can distinguish very well between low (blue) and high vegetation (red). There are problems with correct classification, especially in transition areas and areas in the shade. Image data: Baden-Württemberg State Office for Geoinformation and Rural Development, www.lgl-bw.de, dl-de/by-2-0
Recognition and characterisation of orchards from aerial photographs:
In Johannes Jäger's master's thesis (2025), a segmentation model was trained that is able to recognise orchard stands from aerial images. In combination with available infrared data, it can be used to recognise orchard trees and simultaneously determine the health of the trees (the lower the determined NDVI mean value per tree, the higher the probability that the tree is not healthy).
