Depth feature fusion video super-resolution method based on time sequence grouping

The invention relates to the technical field of video super-resolution reconstruction, and discloses a depth feature fusion video super-resolution method based on time sequence grouping, which solves the problem that reference frame information and complementary information among multiple frames are ignored in the prior art, improves the video super-resolution effect and improves the video super-resolution reconstruction efficiency. A clearer video can be obtained from a low-resolution video through the technology used by the invention. According to the depth feature fusion video super-resolution method based on time sequence grouping, a time sequence grouping mode is adopted, time sequence grouping is carried out by utilizing information of farther frames in an input sequence, so that fine details are recovered by fully utilizing complementary information among multiple frames, the feature fusion performance is enhanced by utilizing farther adjacent frames, and the super-resolution performance of the depth f