Deep Learning-Enhanced ANFIS Classifier for Solar Panel Image Analysis

Abstract

Electricity demand is increasing day by day and hence power utilities are slowly shifting towards renewable energy, mainly solar, as it is more reliable and environment friendly. However, solar power generation systems have very low efficiency and this is the major challenge faced by the researchers. Some of the reasons for the low efficiency is the presence of dust particles, bird droppings, shadows, rain droplets, micro cracks etc. Micro cracks are the major issue to reduce solar panel efficiency. Microcracks are estimated to contribute to a power loss of approximately 80–90%, severely affecting the efficiency and overall performance of solar panels.In this article, the cracked panel and non-cracked panel can be identified by using complex wavelet transform. The Gaussian filter is used to eliminate the distortions in the cracked panel. And this image can be decomposed by sub band images. The corresponding statistical and texture features can be calculated for sub band images and these features are classified using ANFIS classifier. Finally the segmentation algorithm is used to detect the cracked and non-cracked panel images. By comparing with existing methods like Electroluminescence imaging technique, ResNet152 model, Xception model, UAV based thermal imaging technique. The Proposed ANFIS leverages the advantages of both neural networks and fuzzy logic, enhancing the accuracy and adaptability in distinguishing cracked from non-cracked panels. This approach can be deployed in automated inspection systems for large-scale solar farms, enabling early crack detection. By identifying issues sooner, it helps lower maintenance costs while improving the efficiency and longevity of solar panels. Additionally, the method can be integrated with drone-based monitoring systems for remote inspections.