Super-Resolution Reconstruction of Commercial Aircraft Runway Images Based on Omniscient Video Super-Resolution Method

This study focuses on the application of AI in visual navigation during the landing phase of an aircraft, particularly in challenging conditions such as low visibility and night lighting. During the aircraft landing process, the onboard camera's image resolution is reduced due to low visibility and night lighting conditions. This study selects runway simulation images under different airport and weather conditions, such as sunshine, low visibility, and night, to create a super-resolution reconstructed runway dataset containing dozens of runways and a total of 10,000 images. The global Omniscient Video Super-Resolution method is utilized, combined with the hidden state of the front and rear frames, to train and obtain a model applicable to the super-resolution reconstruction of runway images. This innovative model effectively improves the resolution of runway images during aircraft landing.