文档介绍：摘要
摘要
眼睛是人类感官中的重要器官之一,人脑中大约有百分之八十的知识和记忆都是通过眼睛获取的,因此眼睛的健康对于我们来说是非常重要的,而眼底即视网膜是这个器官的重要组成部分,眼底的病变常常与身体其他系统疾病紧密联系,如高血压、糖尿病、心血管疾病、中枢神经系统疾病等。由眼底照相机拍摄的眼底图像是二维的,并且每次拍摄覆盖的眼底区域有限,因此,通常需要在同一患者眼底的不同部位多次成像,即使如此,仍然会不可避免的丢失眼底的三维信息而破坏了眼底图像的完整性。利用眼底图像重建眼底视网膜的三维空间结构,可以方便眼科医生对视网膜进行全方位、多角度的观察,使医生能够更直观、更可靠地检测出视网膜的疾病。
本文在对医学图像处理及三维重建的过程进行了全面深入的分析与研究的基础上,比较了一些算法的优缺点,对基于图像的视网膜三维重建进行了整体设计,包括视网膜图像的增强、配准、层间插值和三维重建。图像增强技术可以消除图像的对比度弱与噪声现象,先采用对数变换增强了图像的对比度,然后采用本文提出的自适应加权中值滤波器来消除图像噪声;由于图像层与层之间会出现错位现象,即平移或旋转等变形, 本文采用基于互信息的图像配准法可以使得不同片层之间的图像得到配准;目前,现有的成像设备所获得的任意两图像的距离均大于每幅图像上各像素之间的距离,直接用于图像的三维重建得到的视网膜三维立体图像会发生畸变,解决的方法就是进行图像的层间插值,本文采用一种改进的基于对应点匹配的自适应优化插值算法插值出中间切片; 最后在上述步骤的基础上采用体绘制算法中的光线投射算法对视网膜图像进行三维重建。
关键词视网膜图像图像预处理图像匹配边缘检测三维重建
Abstract
Abstract
The eyes are one of the most important ans for human eighty percent knowledge and memory in our brains are obtained through the eyes,so the health of the eyes is very important to fundus is an important part of an, the disease in the eyes often associate with other systems of the body, such as hypertension, diabetes, blood diseases, diseases of the central nervous system and so on. The fundus camera can only get the two-dimensional images and its coverage area is limited, so the doctors usually need to take numerous images from different parts of the same patient’s fundus. Even though, it still inevitably lost retinal 3D information which leading to destroy the integrality of the eyes will be convenient for the doctors to observe the full range of retinal by using the retinal fundus images to reconstruct the retina’s three-dimensional space will be also more intuitive and reliable to detect the retinal diseases.
In this paper, prehensively analyse and study the process of medical image processing and 3D reconstruction pare the advantages and disadvantages of various we have designed the retinal image’s 3D reconstruction, including enhancement,registration,interpolation and 3D reconstr