Video surveillance is an important data source of urban computing and intelligence. The low resolution of many existing video\nsurveillance devices affects the efficiency of urban computing and intelligence. Therefore, improving the resolution of video\nsurveillance is one of the important tasks of urban computing and intelligence. In this paper, the resolution of video is improved\nby superresolution reconstruction based on a learning method. Different from the superresolution reconstruction of static\nimages, the superresolution reconstruction of video is characterized by the application of motion information. However, there\nare few studies in this area so far. Aimed at fully exploring motion information to improve the superresolution of video, this\npaper proposes a superresolution reconstruction method based on an efficient subpixel convolutional neural network, where the\noptical flow is introduced in the deep learning network. Fusing the optical flow features between successive frames can\ncompensate for information in frames and generate high-quality superresolution results. In addition, in order to improve the\nsuperresolution, a superpixel convolution layer is added after the deep convolution network. Finally, experimental evaluations\ndemonstrate the satisfying performance of our method compared with previous methods and other deep learning networks; our\nmethod is more efficient.
Loading....