摘要
针对液晶人眼像差校正仪拍摄的小视场视网膜血管图像的拼接算法进行了研究。通过对传统的Harris特征角点算法的改进,解决了该算法在低对比度、小视场视网膜血管图像处理中定位精度较差、需要三次高斯平滑而导致计算量相对较大以及对噪声比较敏感的缺点。然后将配准好的图像进行融合,实现了视网膜血管图像快速、准确的无缝拼接,扩大了视网膜的观测区域,为医务人员的诊治判断提供了更准确的信息。
This paper presented an image stitching algorithm of small view-of-field（VOF） retinal micro blood vessel grabbed with liquid crystal human eye aberration correction system.After improving the traditional Harris characteristic angle method,the following disadvantages were avoided for processing low contrast and small VOF retinal blood vessel： poor positioning precision,cubic Gauss flattening,relatively large calculation quantities,and sensitive to noise.With the presented method,the retinal micro blood vessel can be stitched fast and accurately,which increases the observation field to large extent and provides more accurate information for doctors.
出处
《液晶与显示》
CAS
CSCD
北大核心
2010年第6期884-889,共6页
Chinese Journal of Liquid Crystals and Displays
关键词
视网膜血管
图像拼接
人眼像差校正
retina
image stitching
human eye aberration correction
