Enhanced Detection of Solar Panel Defects Using VGG16-Based Convolutional Neural Networks

Solar energy is a renewable energy source that is expanding rapidly, and it is essential to ensure that solar panel installations are correctly monitored and maintained to ensure that they are operating at their full potential. Detection of various situations and irregularities in solar panel arrays may be accomplished through automated image classification algorithms, which hold great potential. Within the scope of this research, we offer a technique for categorizing images of solar panels that uses the VGG16 convolutional neural network (CNN) architecture. The dataset was taken from Kaggle, which contained 891 Images. The dataset was divided into six categories: Bird drop, Clean, Dusty, electrical damage, physical damage, and snow-covered. In this research, we implemented solar panel image classification using VGG16, a convolutional neural network and achieved a remarkable accuracy of 97.88%. This research exhibited performance compared to the state-of-the-art methods.