No abstract

Ubiquitous applications in diverse fields motivate large-area sampling, super-resolution (SR) and image mosaicing. However, conventional translational sampling has drawbacks including laterally constrained variations between samples. Meanwhile, existing rotational sampling methods do not consider the uniformity of sampling points in Cartesian coordinates, resulting in additional distortion errors in sampled images. We design a novel optimized concentric circular trajectory sampling (OCCTS) method to acquire image information uniformly at fast sampling speeds. The sampling method allows multiple low-resolution images for conventional SR algorithms to be acquired by adding small variations in the angular dimension. Experimental results demonstrate that OCCTS can beat comparable CCTS methods that lack optimized sampling densities by reducing sampling time by more than 11.5% while maintaining 50% distortion error reduction. The SR quality of OCCTS has at least 5.2% fewer distortion errors than the comparable CCTS methods. This paper is the first, to the best of our knowledge, to present an OCCTS method for SR and image mosaicing.

No abstract

No abstract

No abstract