提取鱼眼图像轮廓的算法改进及图像校正

用鱼眼镜头拍摄的图像具有严重的变形,精确地提取鱼眼图像的轮廓是进行图像校正的前提。针对传统的提取鱼眼图像轮廓的算法对噪点抑制能力不强,在有效区域存在大量黑色像素时算法失效等缺点,本文提出了一种改进的提取鱼眼图像轮廓的切线法。实验表明,该算法能准确定位鱼眼图像的圆心坐标和半径,精度高、实时性好,具有明显的优越性。最后,本文用球面坐标定位法对图像进行变形校正,取得了很好的效果。

Improvement of fisheye image contour extraction and image correction algorithm

The use of fisheye lens causes strong distortions in the resulted images.Accurate contour extraction of the fisheye image is a prerequisite for image correction.Traditional algorithm is ineffective to suppress the noises and can not be used when massive black pixels appear in the fisheye image′s circular outline area.So an improved tangent method of fisheye image contour extraction is proposed in this paper.The experiment results show that the proposed algorithm can exact the center and radius of the fisheye images and has obvious superiority of high precision and good real-time.Finally,a satisfactory effect of fisheye image distortion correction is achieved by applying spherical coordinate orientation.