![Institutional Repository [SANDBOX]](images/dias.logo.png "Institutional Repository [SANDBOX]"){kind=logo}
Institutional Repository [SANDBOX]
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/99DAD134-176C-4210-A79F-0CAE10AA5356 | ||
| Year | 2018 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Ioannis Vardaxis, "Design and development of a powerlines fault detection system based on a drone equipped with thermal camera", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2018 https://doi.org/10.26233/heallink.tuc.80051 | ||
| Appears in Collections |
Modern drones use and exploit the vast potential of microprocessor processing power, the tremendous technological advancement in satellites and telecoms to determine the location, the evolution of mobile telephone networks and, finally, the important scientific knowledge and applications in telemetry, image and video processing and transfer, and wireless data transfer. In addition to cameras for image and video recording, the drones can be equipped with thermal cameras to dramatically increase their practical application capabilities. This paper examines the effectiveness of using basic image processing methods and finding their limitations in Matlab in order to detect damage or malfunction to powerlines, with the help of a drone equipped with thermal camera. Three methodologies are proposed, one for locating the cables, one for finding their temperature and one for hot spot detection in the power transmission network. The first two methodologies have in common features like Canny edge detection, application of Hough Transform and creation of areas of interest around each detected cable. In the methodology of finding the cable temperature, two more functions are used, roipoly and sort, in order to more accurately locate the powerlines. In the event of a fault, whether it is a cable cut or a temperature difference between the cables, precise determination of the specific position is made with the help of GPS coordinates. The third methodology proposes a selection of a threshold value for image segmentation using several methods of image processing and analysis. The algorithm is based on finding the max intensity of input image after certain pre-processing procedures. The results have brought satisfying effects and the algorithms, owing to their fast performance, could be used on-line, during vision inspections.
Copyright © DIAS 2013
