| The windage yaw accident of transmission line is one of outstanding issues that often occurs in the power grid, often leading to line trip, windage yaw discharge, wire arc burns,break, etc. It seriously affect the operation of power system, once an accident will lead to huge economic losses.Studies have shown there are many problems about the detection methods of transmission line windage yaw: the method of based on windage yaw computing model needs to build a complex windage yaw model, which cannot be considered the various effects parameters in detail and the parameter error of measurement is larger, computational complexity, low accuracy of wind angle; The method of based on sensor, sensors are generally installed on the ground of insulator string, which causes the wrong judgment easily,and requires the sensor can work under high pressure, strong magnetic fields and other harsh environment for a long time; In addition, the existing methods ignores the flexible of insulator string. there is a shift coefficient difference between the calculated offset and the actual.Based on the above considerations, this paper proposes a detection method of transmission line windage yaw by combination of camera rang and camera calibration,which does image analysis for insulator string, determine the endpoint coordinates of insulators. Finally, we can access to the offset of the real insulator string directly by find out the relationship between the image offset and the actual of insulator string.The insulator string images are obtained by the video camera. To deal with the images by the relevant image preprocessing method, then using the image segmentation method makes the separation of the foreground and the background of insulator string. In addition, according to the morphology of corrosion, the noise and small area is removed. Then according to the mark of the connected area, and the area of the small area is eliminated by the connected area,and the insulator strings are successfully obtained. To extract the skeleton of the insulator strings, then determine the endpoint coordinates on both ends of the insulator string.According to the result of the camera measurement principle and the camera calibration, the realization of insulator string of the deviation of the actual calculation. The main works and innovation points in this paper are as follows:(1) The transmission line windage yaw related theory is in-depth studied,and compare of the advantages and disadvantages for the existing several kinds of polarization detection methods. In this paper, an overall scheme for the on-line monitoring of the windage yaw in the intelligent transmission line is designed, and the main parts of the system hardware are compared and selected.(2) Because of the problem of image caused by the environment, noise, light intensity factors in the video surveillance, the video surveillance image pre-processing method is introduced, which is mainly researched for image enhancement and special occasion image pre-processing method, and then compared to a variety of image processing algorithms to the fog. Finally, the dark channel to fog algorithm is applied.(3) Study area for the insulator string insulator string image feature extraction. Image segmentation, morphological processing, marking regional connectivity methods such as image processing insulator strings, insulator strings extraction area.(4) The detection method of the windage yaw base on the cameras ranging and the making calibration is proposed. Through the camera calibration parameters corresponding experimentally measured cameras, to establish the appropriate model for the partial wind measurements, wind partial derivation calculation expression. Identification out of the skeleton extraction insulator strings, insulator strings successfully coordinate the two endpoints, and then calculate the corresponding partial amount of wind, the wind herein calculated partial amount takes into account the fact that the flexible insulator strings,theoretically calculated results than simply as a rigid body more accurate. |