Evaluation of effect of pre-processing techniques in solar panel fault detection

Solar energy is a clean and renewable source of produced by solar panels. panels deteriorate over time, resulting in generation faults. Faults reduce the overall power capacity photovoltaic (PV) plants. A variety atmospheric functional conditions contribute to formation hotspots on panels, indicating an increase temperature lower efficiency. Early detection faults during PV module inspection monitoring critical for improving efficiency, reliability, safety systems. Thermal imaging non-contact, non-destructive, efficient, effective technique. With thermal image analysis, probable problem areas can be identified fixed before actual failures or problems occur, costs less human labor. Pre-processing one steps automated fault system removing noise artefacts from images. This study investigates impact pre-processing techniques such as filters histogram equalization demarcation accuracy. Five different types faults, single cell, multicell, diode, dust/shadow, PID hotspot are detected. For detection, two segmentation techniques, histogram-based color thresholding RGB channel-based thresholding, applied images Intersection Union (IoU) used determine efficiency techniques. Overall, processed with bilateral filter performed better than other IoU values ranging 0.14 0.5.