A fisheye image barrel distortion correction method of the straight slope constraint

Abstract

Fisheye images always have larger barrel distortion, which seriously prevents feature extraction, target positioning, accurate measurement, etc. Therefore, we present a fisheye image barrel distortion correction method based on the straight slope constraint in the paper. Compared with other methods based on a single image, the method not only eliminates the residual errors on the edge of the image and asymmetry, but also reduces the difficulty of calculation parameters. The basic idea is shown as follows: the constraint condition that the slope is equal on the same straight is utilized for replacing the lens projection process. Then, distortion parameters are calculated by a single image which is captured in front of the grid target. In the paper, a method of segmenting grid lines of the high dynamic image is proposed so that the center of the lines can be accurately extracted. In addition, we also put a method of accurately calculating the distortion center by the fundamental matrix in the paper. The method is separated from the process of calculating distortion coefficients for removing coupling errors. Experiments show that the method in the paper has ideal correction effects, and the correction precision is 0.42pixel.

The Automatic Undistortion Strength Estimation for Any Describable Optical Distortion

“…1. For all the testing images the ∈ [1,3] and = 0.02 , which give 100 images for processing. The first column is the best correction parameter estimation.…”

“…The recent research [3] deals with the fisheye barrel distortion effect. The correction method uses the benefit of the straight slope constraint.…”