Usage of GA with Multilevel Thresholding to Detect Ice Thickness of Iced Conductor

Abstract

Ice load on the electrical transmission line (ETL) can change the aerodynamics of the lines, causing galloping and faults, for example short circuits. If enough ice load forms, the weight of ice on the line can cause the electrical lines to collapse and then it can cause loss of load. Thus electrical energy of some consumers may be cut for days. So, ice load must be monitored continuously to prevent this case. Image processing can be used to monitor ice load and to determine ice thickness of iced conductor. Ice thickness of iced conductor can be determined by using image segmentation, and image segmentation makes according to optimum threshold value. Optimum threshold value can be determined bi-level threshold method or multi-level threshold method. It was seen in literature that multilevel threshold method is an effective method in object recognition. Multilevel thresholding can be made by Otsu method. But determination of optimum threshold level is difficult process. In this study, multilevel threshold method was used and its optimum threshold level was determined with genetic algorithm. Objective function of genetic algorithm is determined by Otsu method.