UAV Image Denoising based on Swin Deformer for Power System Condition Monitoring
Dong Wang, Linwei Ye · 2023
In order to improve the quality of UAV transmission line inspection pictures and ensure the effectiveness of the defect recognition network, this paper proposes a model based on Swin Deformer to denoise the pictures. The data set of this method is the original data set collected by UAV, and the test set and training set are obtained by classifying the raw data. At the same time, Gaussian noise is added to generate tags to ensure the effect of supervised learning. Finally, the test set is input into the pre-training model, and the results obtained prove that our noise reduction network can greatly improve the image quality and further improve the effect of line maintenance.