LIRNet: A Lightweight Inception Residual Convolutional Network for Solar Panel Defect Classification

Solar-cell panels use sunlight as a source of energy to generate electricity. However, the performances solar decline when they degrade, owing defects. Some common defects in solar-cell include hot spots, cracking, and dust. Hence, it is important efficiently detect repair them. In this study, we propose lightweight inception residual convolutional network (LIRNet) panels. LIRNet neural model that utilizes deep learning techniques. To achieve high performance on panels, including fault detection accuracy processing speed, draws hierarchical learning, which two-phase solar-panel-defect classification method. The first phase data-preprocessing stage. We K-means clustering algorithm refine dataset. second training model. designed powerful enhance speed up time. experiment, improved by approximately 8% performed ten times faster than EfficientNet.