An improved longitude-latitude mapping algorithm for fisheye image calibration

Field of view (FOV) of the fisheye lens is close to or even higher than 180°, which brings about extraordinary imaging effect that the visual range would be far wider than the human eye perspective. However, the wide visual range of the fisheye lens is always at the expense of image quantity, so that the images obtained by fisheye lens would inevitably show a large degree of distortion aberration. In this paper, we propose an improved distortion calibration algorithm of longitude-latitude mapping to eliminate the image distortion aberration caused by fisheye lens. This method does not need to depend on a specific physical camera and has general and universal significance. By using Matlab as a mathematical analysis tool, the verified experiment and the corresponding write code is performed to correct a specific fisheye lens’ image based on the previously proposed algorithm. A reverse mapping is used to avoid cross-border problems, and comparative experiments are also analyzed to show the difference between the image processing with the traditional and the proposed algorithm. Experimental results demonstrate that the proposed algorithm can better achieve the correction of fisheye lens distortion. The twisted lines that are not expected can be well corrected into the straight lines. This method can not only reduce the problem of horizontal stretching of the image, but also make the processing results visually consistent with people's viewing habits.