The Method Of Image Recognition For Broken Strands Of Transmission Line And Self-shattering Of Insulator

Transmission lines act as a very important role of Transporting electricity across country. As the Transmission lines are located in the harsh natural environment, some malfunctions often occur such as Broken strands of Transmission line、Self-shattering of insulator,etc. Because these defects will cause great harm on Transmission lines, they must be identified and repairs in time. Traditional "transmission line" approach mainly rely on manual Monitoring, which has low working efficiency and work intensity. With the development of computer vision, image processing and communication technology,a new transmission line approach – “Remote Intelligent monitoring transmission line”came into being. The approach convey the HD images to the center of monitoring transmission through the Communications network, then the transmission line defects will be detected using the technologies of image processing. This new approach will greatly reduce the work intensity while improving efficiency.In this paper, XiChang-DongJin remote intelligent transmission line systems act as an engineering background. Selecting the wire and insulators in the transmission line as a research target and researching image recognition algorithms, which can find Self-shattering of insulator and Broken strands of Transmission line. The main work of this paper includes the following aspects:1. Introducing the research status of transmission line defect identification and proposing research objectives, then introducing image processing key technologies which will be used in the follow algorithms.2. Designing the image recognition algorithms which can find Self-shattering of insulator and Broken strands of Transmission line in the natural environment. As for the insulator, at first, using the feature in Lab space and the connected domain features to segment the insulator from the natural environment, then creating a mathematical model according the segmentation insulator. The defects will be found through the region feature. As for the wires, using domain connectivity features and Hough line detection to extract the wire. Then define the concept of “Degree of dispersion”. According to the concept to determining whether the wire has defects. The tests show that the algorithm can adapt to the complex natural environment, and can pinpoint the exact location of the defect occurred.3. Introducing the intelligent remote power transmission line platform, On this platform recognition rate was tested. Meanwhile, in order to improve the recognition rate, the failure cases were analyzed and some measures were raised to correct the failure case.