89 Drone Based Aerial Inspection of Energy Systems H. Petersen and O.L. Christiansen DTU Fotonik, Technical University of Denmark ABSTRACT A national goal for Denmark is to fully cover the national energy consumption by renewable energy sources by 2050. A large part of the national energy consumption is scheduled to be covered with the use of solar energy, which means that the amount of solar power plants is increasing rapidly.1 Continuous maintenance is important in order to ensure the optimal operation and longevity of energy systems. Manual maintenance of large energy systems such as solar power plants, district heating, overhead power lines etc. can be complicated, time consuming and energy inefficient. As an example the traditional method of inspecting high voltage cables is using a helicopter.2 A preferable low cost and energy efficient method of inspection could be implemented using an unmanned aerial vehicle (UAV) equipped with an infrared camera. Using an infrared camera it is possible to detect faults in the systems, which can be characterized by hot and/or cold spots. As the amount of images needed to cover an energy system is proportional to the size of the system, an effective image coding scheme is needed, to accommodate limited storage capabilities on board the UAV and/or limited bandwidth for real-time inspection via a wireless video link, while maintaining an image quality high enough for detection of defects. Furthermore, the future development of infrared cameras of higher resolution, will increase the need for effective image/video coding techniques handling the specific characteristics of infrared images/video such as a bit depth of 14-bits compared to the 8-bits used per colour component in standard digital images of the visual light spectrum. The main aim of the project is to study the application of the JPEG2000 image codec, for the coding of infrared images obtained using FLiR Tau and FLiR Quark cameras. JPEG2000 is an extension of the standard JPEG codec, commonly used for the compression of the images in the visual light spectrum. JPEG2000 supports encoding of 14-bit images and incorporates wavelet theory to improve image quality at lower bitrates. The image/video coding will be tested using a UAV at a solar power plant located at DTU Risø in Roskilde. The results of several test flights will be presented at GRØNDYST. ACKNOWLEDGEMENTS Supervisors: S. Forchhammer and C. Mantel REFERENCES 1 Denmark reaches 2020-goal for solar energy before time. (2012, September 12). Retrieved May 18, 2016, from http://um.dk/en/news/newsdisplaypage/?newsid=25147b44-3dce-4647- 8788-ad9243c22df2 2 Helicopters used to conduct transmission line inspections in Toledo Edison services area. (March 2, 2016). Retrieved May 18, 2016, from http://www.pennenergy.com/articles/pennenergy/2016/03/helicopters-used-to-conducttransmission line-inspections-in-toledo-edison-services-area.html 404 COMMUNICATION POSTER IDEA BACHELOR FINAL ASSIGNMENT
Grøn Dyst 2016
To see the actual publication please follow the link above