Study On Measuring Transmission Conductor Sag Based On Image Processing Algorithms

In recent years, along with the development of UHV transmission project, the chargeload of transmission lines is aggravated, and hence the allowable temperature of transmissionlines has increased from70oCto80oC, which will also increase conductor sag of thetransmission lines. In southern China, the accidents caused by icing and snowing of overheadtransmission line occur frequently. The icing conductor increase wire weight as well as theconductor sag, further affected by the wind, it is so easy to happen conductor galloping underthe action of wind force. Conductor sag is one of the key factors to the safe and reliableoperation of transmission line; therefore, how to realize the measurement of conductor sag intransmission lines became the focus of research. The existing conductor sag measurementmethods including the indirect measurement of the conductor stress or the conductortemperature to determine conductor sag and by using total station instrument without prism tomeasure conductor sag. However, the intelligent image processing is used to measure theconductor sag in this paper, its convenient measuring operation, high precision, intelligenceand other advantages all accord with requirements for the state detection of transmission linesin smart grid.Through the analysis of the live images of transmission lines, it was found that thelinearity, spectral characteristic of conductor in transmission lines and the angular pointcharacteristic existed in conductor interval bar, coupled with that the fog and cameramovement will often lead to image blurring since transmission line is often located in the field.According to the above factors of the live pictures in transmission lines, different filter is usedto do image filtering and image enhancement processing, then after the image processingresults are analyzed, three method for conductor feature extraction are proposed, they areseparately based on conductor edge feature extraction, gray level of the conductor and theimproved Hough transform, meanwhile, the three methods are all validated, by thecomparison and analysis, it is found that the effect of the method based on the combination ofSUSAN and Harris for conductor feature point extraction is preferable under the poor lightingcondition.At first, carring on the camera calibration on the basis of the camera’s imaging model.Secondly, according to the binocular stereo vision technology, carring on the epipolarrectification, stereo matching and3D point recovery. Finally the3D information of conductoris restored by stereo matching on feature points. In order to facilitate curve fitting ofconductor, the reference frame of the feature points must be transformed to the reference frame that use the left suspension point as origin, after coordinate transformation, it will beeasy to do the two-dimensional curve fitting for conductor feature points, and then it comesthe calculation of the lowest point of the conductor sag, lastly is the comparison between themeasured value and the actual value of the conductor sag, the comparison result shows thatthe average relative error reached4.095%, which illustrates the feasibility of the algorithm.