Improving Defect Classification in Photovoltaic Modules Using CNN with Attention-Based Processing

This paper presents the development of an advanced convolutional neural network (CNN) model that incorporates attention-based processing to enhance defect classification in electroluminescence images. The attention mechanism enables the model to focus on significant regions of the image, thereby improving its ability to capture critical features. By selectively highlighting these regions, the model achieves greater accuracy in identifying subtle defects that may otherwise be overlooked. Unlike traditional CNN approaches, which rely solely on convolutional layers, the proposed model leverages attention to dynamically concentrate on areas of interest, leading to more precise defect classification. The proposed method demonstrates an accuracy of 93%, which represents a substantial improvement over conventional techniques in detecting and classifying diverse defect types in electroluminescence images.