立体断层图像配准误差校正仿真研究

摘要

In this paper,we propose an error correction method for three-dimensional sectional image registration based on low-rank texture extraction.The collected background image and feature image of registration error are differenced to form texture space on image surface,and TILT method is used to divide image into various sub regions,and then Jacobian matrix of image registration error is formed and correction threshold of registration error is given.Finally,threshold is used to correct registration error in three-dimensional sectional image transmission.Simulation results show that the proposed method has high correction accuracy of registration error.It retains the texture information of three-dimensional sectional image in a comprehensive way.%对断层图像的配准误差进行校正,能够有效解决断层图像传输过程质量降低问题.对断层图像配准误差的校正,需要组建图像配准误差雅阁比矩阵,给出配准误差校正阈值,完成图像传输过程配准误差的校正.传统方法定义断层图像纹理联合关联规则,引入兴趣度阈值衡量图像纹理的偏差,但忽略了得到配准误差校正阈值,导致校正精度偏低.提出基于低秩纹理提取的三维断层图像配准误差校正方法.将采集到的背景图与配准误差特征图进行差分,构造图像表面的纹理空间,利用TILT方法将图像分割为多个子区域,组建图像配准误差雅阁比矩阵,给出配准误差校正阈值,利用该阈值进行三维断层图像传输中配准误差的校正.仿真证明,所提方法配准误差校正精度高,较为全面地保留了三维断层图像的纹理信息.