Distortion Correction of Building-Fisheye-Images

The fisheye image is the image taken with a fisheye camera, in which the object suffer from distortion. The street view building images are always taken by fisheye cameras due to the boarder view of the very cameras. So, the distortion of the images must be corrected before further applications. This paper compared three state of the art's distortion correction algorithms, including algorithm based on Polynomial Model, algorithm based on Division Model, and algorithm in [1]. The comparison results on both synthetic and real data show that algorithm in [1] performs the best.