Application of Convolutional Neural Network Based on Attention Mechanism in Surface Defect Recognition of Power Equipment

With the continuous expansion of power systems, traditional manual inspection methods are increasingly unable to meet the demands for accuracy and efficiency. In recent years, image-based intelligent recognition approaches have emerged as promising techniques for detecting surface defects in power equipment. This paper proposes a convolutional neural network integrated with a convolutional block attention module (CNN-CBAM) to identify surface defects in power inspection images. Experiments were conducted on a dataset comprising both synthetically generated and partially real images, covering five typical types of surface defects with a total of 3,000 samples. The results show that the proposed model outperforms conventional CNN and ResNet50 architectures in terms of accuracy, recall, and F1-score, demonstrating its effectiveness and practical value for surface defect recognition in power equipment.