Detection of solar panel defects based on separable convolution and convolutional block attention module

The share of renewable energy in the electricity market is increasing year by year. It is necessary to identify damage of solar panels in a timely manner, as solar panels are important components in photovoltaic power generation. In this paper, a lightweight solar panel fault diagnosis system based on image pre-processing and an improved VGG-19 network is proposed to address the problem of blurred solar panel field images, which are not easy for defects detection. First, we use Daubechies 4(DB4) wavelet and morphology-based enhancements to improve the quality of solar panel images. Then, conventional convolutional layers in the VGG-19 are replaced with separable convolutional layers to reduce the number of network parameters and improve training efficiency. Finally, the Convolutional Block Attention Module (CBAM) is introduced to improve the accuracy of solar panel defects' detection. A dataset consisting of 3344 images of solar panels was used to evaluate the performance of the proposed method in defect detection. The experimental results show that the method has an accuracy of 87.8% and a detection speed of 0.047 s per image. The proposed model has higher accuracy and more stable performance than other conventional networks with a lightweight structure, demonstrating the reliability of the improved VGG-19 in detecting solar panel defects in practical applications.