Research On Detection Method Of Ice Coating Thickness In Transmission Lines Based On Image Processing

Overhead transmission lines were an important part of the power system and the basis for the safe operation of the power grid.The ice coating of transmission lines had become an important factor that be harmful to the safe and stable operation of the lines.Severe ice-covering could lead to wire galloping,flash-over of insulators,and even accidents such as disconnection and tower collapse.This caused inconveniences to people's production and life,and also brought huge losses to social property.Therefore,it was of far-reaching significance to research a method to monitor the ice-covering conditions and thickness of transmission lines to ensure the safe operation of the power grid.Aiming at the deficiencies in the existing ice detection methods for transmission lines,this paper proposed a detection method for transmission line ice detection based on image processing.Its core idea is that after the wire was covered with ice,its radial width was increased compared with that before uncovered ice,and has important significance for deicing the line.The degree of increase reflected the thickness of ice.The accuracy of the ice coating thickness directly affects the setting of the melting ice current and time of the ice melting device.After the existing image-based transmission line ice coating thickness detection methods was analyzed,the key technologies of the three methods,image denoising,edge detection,and ice overlay thickness calculation model,were analyzed and improved.In the aspect of image denoising,the conventional denoising method made the edge information blurred,which was not conducive to the disadvantages of subsequent edge detection,and proposed an improved image denoising algorithm for compressive sensing.Firstly,the denoising reconstruction model based on compressed sensing was analyzed and modeled.Then,the denoising model was improved by using the maximal function to smooth the lp norm,and then the optimal solution of minimum lp norm was approximated by discretizing the solution sequence of the smooth function.Experiments show that this method not only had a good denoising effect on the image,but also could preserve the integrity of the edge information of the image,which provided the conditions for the accuracy of subsequent edge detection.In the aspect of edge detection,firstly through the principle of edge detection of Canny operator,it was found that the automatic and accurate acquisition of high and low thresholds was the key to adaptive edge detection when applying Canny operator for edge detection.Subsequently,an improved edge detection method for the Canny operator was proposed.The optimal threshold gray-level segmentation method was used to obtain the high threshold of the Canny operator,and then the high-low threshold correspondence was used to determine the low threshold.By setting the target area,the complexity background could be removed.The complete ice-covered wire edge would be obtained and the influence of the pseudo-edge could be eliminated.In the aspect of ice thickness calculation model,a segmented thickness calculation model was proposed.First,the overhead transmission line has a certain overhang arc was considered,the transmission line in the image was divided into equal length sections,so that it could be regarded as a straight line,and the average width between the upper and lower contours of the transmission line in each small section was easy to find.Then,in order to avoid the large error of the overall result due to the local error being too large,some segments whose calculated values seriously deviate from the overall average value were ignored.Finally,the average value of the average width of a number of transmission lines is calculated,that was the measured width of the transmission line pixel.When the ice thickness of the transmission line is different,the comparison experiment with the traditional detection method is conducted,the result showed that the ice coating thickness and the practical measurement value were very small,which verifies the effectiveness and feasibility of this method...