摘要
提出鱼眼图像轮廓提取的扫描线逼近算法,算法能够准确地求取鱼眼图像的圆心坐标和半径,特别是在鱼眼图像圆形轮廓区域内存在大量黑色像素点、统计算法失效的情况下,该算法仍能准确地提取图像轮廓,计算图像定位参数,具有明显的优越性。在此基础上,采用球面坐标定位校正方法进行图像扭曲变形校正,取得了较好的图像校正效果。
Scan line approach algorithm is put forward to calculate the center coordinate and the radius of the fisheye image circle, which particularly has the obvious superiority in the condition that massive black pixels appear in the fisheye image circular outline area and the traditional statistical algorithm fails, And a satisfactory effect of fisheye image distortion correction is ultimately achieved by applying spherical coordinate orientation approach based on scans line approach.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第12期2878-2879,2882,共3页
Computer Engineering and Design
基金
国防科技预研基金项目(51404040204JB1105)
关键词
鱼眼图像
轮廓提取
扫描线逼近
球面坐标定位
变形校正
fisheye image
edge linking
scan line approach
spherical coordinate orientation
distortion correction
