Electric Power Icing Detection Based On Hypercomplex Frequency Domain

Recently, with the frequent occurrence of power accidents such as power cut in large area, tower collapse, insulator damage and so on, which are caused by the the disaster of overhead electricity transmission line covered with ice and snow. Therefore, it plays a vital role in the normal operation of State Grid and people’s convenient life whether the overhead electricity transmission line and the area covered with ice and snow can be detected accurately. In order to accurately detect the icing areas which exist in an image of electricity transmission lines and power lines, so that calculate the ice thickness, a direction controllable electricity transmission line detection algorithm which can adjust to the object direction in an image and a edge extraction of electricity transmission line covered with ice image algorithm is presented urgently needed.In this paper, to solve the problem brought by the difference of electricity transmission line directions, complexity of background and the inaccurate of image edge extraction, we proposed an electric power ice detection methods based on hypercomplex frequency. The main work is shown as follows*:In the image edge detection aspect:we propose a new image edge detection algorithm based on hypercomplex frequency which combines the Fourier Transform of2-dimensional2-order and2-dimensional4-order steerable filter and hypercomplex. The proposed algorithm can do selective detection for different directions in an image while filtering out the edge information in the other directions. We also proposed a new idea: take three imaginary sections i、j、 k to simultaneous detect three color channels RGB of natural scene separately, and then we can obtain detection results through calculating the detected RGB three channels image. Experimental results show that compared to traditional image edge detection algorithms, our algorithms can detect target object more accurately in selected direction.In the detection of electricity transmission line and the area covered with ice aspect:we proposed an optimal direction ridge energy extraction algorithm based on frequency domain hypercomplex algorithm. First, detect edge information of electricity transmission lines covered with ice in the selected direction through2-dimensional2-order hypercomplex steerable filter, then detect ridge point information through orthogonal hypercomplex filter, and then obtain the ridge energy function of electricity transmission line covered with ice according to direction information and ridge point information. In the final, we can obtain optimal ridge energy detection of electricity transmission line covered with ice through the ridge energy function. The proposed algorithm can take advantage of the three color channels of the image information to accurately detect areas of electricity transmission line covered with ice.Finally, we build a visualization Human Machine Interface (HMI) platform for detection of power transmission line covered with ice using MATLAB_GUI graphical user interface software.