SUPERRESOLUTION ALGORITHM BASED ON SINGLE-FRAME MOTION FIELD

Many computer, machine or robot vision applications need increased resolution of images. Very actual are security applications and vision for mobile automatic vehicles and robots. The application of superresolution is interesting in airborne, land and maritime applications. Adverse illumination and noise image are particular challenge. Superresolution algorithm based on motion field is developed in the article. It is applicable on the single image. For real time applications, the time of execution ranges from 7.6 fps up to 34 fps in the experimental setup (Matlab, Windows, etc). There is still a lot of place for improvements and further research.