This reprint aims to immerse the reader in the latest surface defect detection methods of industrial automation. From distribution-preserving under-sampling to semi-supervised inspection, with the use of swin transformer techniques, we aim to elucidate the application potential of these advanced technologies. In this reprint, research experts in the field will publish their results and prospects for the future. These research results will not only improve detection accuracy but also greatly accelerate detection speed, enhancing the process of industrial automation. We will provide more advanced and accurate detection technology for product quality inspection so that artificial intelligence technology can be integrated into and further developed in the industrial field.