Deblurring Reconstruction of Monitoring Video in Smart Grid Based on Depth-wise Separable Convolutional Neural Network

Abstract

To improve the definition of monitoring video in smart grid, a de-blurring reconstruction model of monitoring video in smart grid based on depth-wise separable convolutional neural network is proposed. Firstly, a multi-scale feature extraction convolutional neural network is used to extract rich feature information of the input video frames. Secondly, all support frames are aligned to the target frame from the two ends of the input video sequence in chronological order, and a time and space self-attention mechanism is used to fuse the support frame features to the target frame. Thirdly, an improved depth-wise separable residual network and pixel shuffle up-sampling network are constructed to perform high-definition reconstruction of the target frame after feature fusion. Finally, the model is trained with public datasets and video datasets of smart grid and is applied to the video de-blurring reconstruction of monitoring system of smart grid. The test results indicate that the proposed model can effectively improve the definition and visual effects of monitoring video, whose average values of the peak signal-to-noise ratio and structural similarity index reach 32.18dB and 0.9132 respectively.