Research on Video Super Resolution Reconstruction Method Based on Sparse Camera Array

How to obtain high-resolution and high-definition images through sensors is of great significance in scientific research. When the camera optical structure and image sensor are constrained, super-resolution reconstruction is an effective method to improve image resolution. By utilizing information from images and videos, super-resolution reconstruction can obtain higher resolution images without changing the inherent imaging performance of the system. Based on this background, we propose a video super-resolution reconstruction method with long-term feature extraction. By introducing iterative and cyclic propagation mechanisms, the information in the overall image sequence is fully utilized to perform long-term feature extraction and super-resolution reconstruction on low resolution videos. An algorithm is designed to use sparse camera array to capture multiple video sequences for super-resolution reconstruction. To guide feature matching, alignment, and distortion between different cameras, we introduce the optical flow mechanism, and the redundant information of support frame cameras from different perspectives is fully utilized to perform super-resolution reconstruction on videos captured by reference frame cameras. The proposed method has improved both peak signal-to-noise ratio and structural similarity index compared to traditional methods.