We are pleased to announce that Marcel Neuhausen successfully passed his disputation on Monday. His first examiner Prof. Dr.-Ing. Markus König as well as his second examiner Prof. Dr. Christian Koch came to a positive result!

His doctoral thesis is titled: "Assessment of Computer Vision Methods in Civil Engineering". The thesis deals with the use of computer vision in civil engineering and the development and evaluation of image processing approaches.

Traditional deep learning methods are up for discourse due to high performance requirements and ease of use. Methods such as Soft Cascade and Yolo (You Only Look Once) are therefore available as promising alternative machine learning approaches.
In this dissertation, image processing approaches have been developed which are based on these methods. These were evaluated in different scenarios and compared with respect to the different requirements that the construction industry entails.
Another difficulty is the availability of civil engineering tasks, which is why synthetic data were considered for training the machine learning methods as well as for evaluation.

The results show that both methods have a low barrier to entry, provide reliable object recognition and are suitable for tracking. However, there are also different characteristics, so that the specific application must be taken into account. Furthermore, it was shown that by enriching the training data with synthetic data, a significant improvement in recognition performance can be achieved. This comparability was also confirmed in the evaluation. Consequently, the use of synthetic data can support a comprehensive evaluation of computer vision approaches in civil engineering when data is scarce.

We all congratulate you very much for your great work and wish you all the best!

•    Marcel Neuhausen, M. Sc. (more…)
•    Markus König, Prof. Dr.-Ing. (more...)