Insulator condition using local binary patterns combined with support vector machines

Insulator is an important component in the power grid. Therefore, faulty insulator can cause a great damage to the power grid that would lead to leakage currents flowing through line supports. This leads to increase in electrical loses, voltage drop and put human safety to risk. Hence, it is very important to monitor the condition of an insulator before resulting to a great damage in the power grid. Computer vision is recognized as a means to solve this problem safely, speedily and accurately instead of the manual method of monitoring. This paper presents insulator condition using local binary patterns combined with support vector machines. Experiments show an improved performance when local binary patterns is used as a feature extraction method over gray level co-occurrence matrix combined with support vector machines. Results obtained are presented and discussed.