Nonstationary fuzzy set based image mosacing approach in cyber-physical systems

The usage of cyber-physical systems is increasing day by day thanks to the advantages of it provides. But, like in every system, uncertainties are threat factor for cyber-physical systems which are used in large-scale and mission-critical applications generally. One of the biggest problem in handling uncertainty is that they cannot be modelled in a deterministic way. That is why condition monitoring software is used for minimizing the harm which unexpected situations like uncertainties cause. Monitoring of images taken through camera by an operator or processing the images by image processing algorithms are ways that are used in condition monitoring. But, it makes hard to monitor the system in a frame in the case of being the system in a large area or in indoor. In this thesis study, image mosaicing approach is proposed for monitoring of large-scale cyber-physical systems which have uncertainties. Feature-based image mosaicing methods basically consist of detection of interest points, feature extraction, image matching, homography estimation, transform and color blending steps. The proposed feature-based image mosaicing approach uses nonstationary fuzzy logic in the feature extraction step which is the most important step in image mosaicing. Using nonstationary fuzzy logic in this step improves the success of the approach, and it causes providing good results especially for contaminated images. In order to test the image mosaicing approach, a cyber-physical system is realized, and images are obtained from this system. As a result, in this thesis study, a method based on nonstationary fuzzy logic is developed to stitch the images taken from sample cyber-physical system, and it is verified with the tests applied. This thesis study is supported by FÜBAP, Project number: MF.17.09, and various national or international publications about the thesis studies are published.