摘要
对基于联合变换相关器的像移测量数学原理进行了说明,介绍了面阵CCD的安装位置以及输入图像的获取方法,分析了低信噪比输入图像、相对转动位移对测量精度的影响。应用MATLAB对输入图像进行傅里叶变换得到联合功率谱,对二值化处理的功率谱进行傅里叶变换得到相关输出,用质量中心算法计算出像移量。分析得出,输入图像信噪比大于5dB,测量均方误差小于0.05像素;输入图像相对转动位移在0.1°的范围内,转动位移对测量精度和峰噪比的影响可以忽略不计。综合分析表明像移测量均方误差小于0.05像素。
The mathematic principles of image motion measurement with joint transform correlator are clarified. The position of matrix CCD on the focal plane, the method of getting input image are presented. The effect on the accuracy of lower SNR input image with angle displace to the measurement is analyzed. The input image is Fourier transformed to get the spectrum image. The spectrum image is binaried and Fourier transformed to obtain the correlation image. The corresponding shift value is calculated with centre of mass calculation. The results show that, when the SNR of the input image above 10dB, the RMS of accuracy is less than 0.05 pixel, and when the angle displace within 0.2°, the effect on accuracy and PNR can be dismissed. The simulated accuracy is less than 0.05 pixel when all factors took into account.
出处
《光学技术》
CAS
CSCD
北大核心
2008年第5期781-784,共4页
Optical Technique
关键词
联合变换相关器
像移测量
傅里叶变换
质量中心算法
joint transform correlator
image motion measurement
Fourier Transform
centre of mass calculation