Study On Power Lines Corridor Features Extraction Method From Different Remote Sensing Data

The rapid development of the world power grid brings up higher requirement to the security maintenance of electricity grid and other infrastructure facilities。Over the past years, the power lines inspection and maintenance mainly depend on manual survey and construction, which is inefficient, high cost and poor security。For this reason, it is an urgent requirement for power enterprise to make use of remote sensing technology for power line inspection. When you prepared to establish the real-time 3D power lines inspection geographic information systems or accurately measure the distance between the power lines and corridors vegetations, the identification of type of remote sensing data and corridor feature extraction method should be considered. This is also the research objective of this dissertation。In this thesis, the corridor feature extraction method based on different remote sensing data is first studied. And the application advantages of remote sensing data for power lines inspection are also discussed. The concrete research content includes the power lines automatic extraction method from UAV aerial images, the accuracy segmentation and height estimation method of tree canopy based on multi-spectral images, power lines extraction and 3D construction method from airborne LIDAR data. At last, an economic, rational, efficient remote sensing data acquisition program is also discussed. The main innovation results of this dissertation are listed as follows:(1) The power lines extraction method from high resolution optical UAV RS image are studied in detail. The properties of power lines in aerial optical images are firstly analyzed and summarized, then the advantages of different edge detection operator for straight line detection are reviewed and compared. A strong anti-noise Gabor filter operator is presented and the parameter selection of Gabor filter is also in depth analyzed. Finally an improved Hough transform for power line segments fitting is introduced. This experiment shows that the Gabor filter is obviously better than other edge detection operator in line detection and background noise removal.(2) The method of corridor tree crown extraction and height calculation based on high-resolution multi-spectral NIR image is discussed. First of all, the DSM and DOM of power lines corridor are obtained using the photogrammetry technology, and then the tree crown areas are fast, accurately achieved by calculating the NDVI vegetation index, and finally the tree crown height is estimated through the detected crown area and DSM. In this paper, the tree crown area detection method base on calculating vegetation index is more accuracy than the conventional extraction method using edge detection. In the part of tree crown height estimate, the image geometric distortion and ground fluctuation are both considered. So the canopy extraction and elevation estimation method is more precise and perfect.(3) This dissertation has made a useful exploration for power lines automatic extraction and 3D reconstruction from airborne LIDAR data. This paper has made an earlier study on the automatic power lines extraction and fitting method. Firstly, the characteristics of power lines in LIDAR data are reviewed, and then a kind of cloud points filter based on angle and a power line points extraction and automatic 3D reconstruction procedure are designed. The ability to extract the major power lines points from LIDAR data and automatic curve fitting to form a precise, continuous power lines corridor is the characteristics in this method.On the above basis, the merits and prospects of using different remote sensing data for power lines corridor inspection were in a brief summarized. For the research achievement, there is some reference value and good application prospect on power lines security inspection and the construction of 3D power lines inspection geographical information system.