摘要
根据互相关因子算法和绝对差分算法的基本原理,利用Matlab实现了倾斜信息提取。在此基础上,考察了两种算法的精确性问题,发现绝对差分算法的拟合精度高于互相关算法,其数量级可达到0.0001像素;结合哈特曼波前传感器的测量结果,采用比对的方法,对有结构扩展目标两种算法提取倾斜信息的有效性进行了验证,可以得到,不同倾斜范围时两种算法提取的信息高度吻合,且处理后与哈特曼波前传感器走势一致;经自适应校正后,得到了很好的长曝光像。实验结果表明,这两种算法均可有效地提取出自适应光学所需要的光束波前整体倾斜信息。
According to the basic principle of cross-correlation algorithm and the sum of absolute difference algorithm, the tilt information is extracted by Matlab. The accuracy of above two algorithms is investigated on this basis. It is discovered that the fitting accuracy 0.0001pixels of the sum of absolute difference algorithm is higher than that of cross-correlation algorithm. Combining with the measured results of Hartmann-Shack wavefront sensor, the validity for structural extended object with two algorithms and the tend of information in the different tilt range is consistent with the H-S wavefront sensor; Applying two algorithms to adaptive correction we obtain better long exposure image. The experimental results show that the global tilt information of beam wavefront for adaptive optics can be effectively extracted by both algorithms.
出处
《光电工程》
CAS
CSCD
北大核心
2004年第11期5-8,共4页
Opto-Electronic Engineering
基金
国防重点预研项目
关键词
自适应光学
扩展目标
信息提取
整体倾斜
互相关因子算法
绝对差分算法
Adaptive optics
Extended object
Information extracting
Global tilt
Cross-correlation coefficient algorithm
Absolute difference algorithm
