Detection of pollution using image processing techniques in photovoltaic panel surface

There are lots of factors which affect the efficiency of energy produced in solar power plants. Solar radiation, speed and direction of the wind, photovoltaic cell type, positioning of PV panels, installation type, temperature and humidity of the installed region, dust and snow on the panels are the most important factors affecting the Efficiency. In this study, the efficiency effect of the dust accumulated on the panels was studied. Firstly, the panel was dusted in three different dirt levels, which were Little Dirty (LD_Naturel), Normal Dirty (ND_3gr/0.388m^2) and Very Dirty (VD_9gr/0.388m^2). Then, some system data such as PV panel current, voltage, irradiation value and PV panel cell temprature were obtained from dusty panels. Artificial solar source with halogen lamps were used so as to get data about PV panel system. In addition, electronic hardware and software system were developed in order to observe realtime intantenous power output depending on dirt level. Hence, the effect of the dust factor on PV panel efficiency was interpreted by confıguring with numeric values and instantenous power outputs depending on dirt level were compared. Besides, some images were obtained via a camera with Internet Protocol (IP) from PV panels periodically for the detection of dirt level. By these PV panel dirt images, 19 Feature Numbers (FN) were obtained based on Grey Level Co-Occurance Matrice. PV panel's dirt level was classified by using these new featured data based on Artificial Intelligence (AI). Consequently, the detection of PV panel dirt level will give an idea about the time and frequency of PV panel cleaning to the personel users and Solar Power Plant (SPP) operators using PV panels.