摘要
为了测量由卫星姿态不稳定或振动等原因引起的空间相机亚像元像移,对基于联合变换相关器的像移测量方法进行了研究,并进行计算机仿真。阐述了该方法的数学原理,面阵CCD的安装位置以及输入图像序列的获取方法。应用MatLab对图像进行线性插值获得具有亚像元像移的图像,并添加噪声使图像具有较低信噪比;对图片进行傅立叶变换得到联合功率谱,二值化处理后的功率谱再次进行傅立叶变换得到相关输出;用质量中心算法计算出像移量。结果表明,像移测量的精度不随景物的内容而变化,测量得到的像移误差均方根误差值小于0.2个像元。
To measure the sub - pixel image motion which is caused by satellite attitude instability or vibration, the joint transform correlator method was used and simulated. The mathematics principles were clarified, the position of matrix CCD on the focal plane and the method of getting input image were presented. Based on MatLab, sub - pixel image motion was oblained by linear interpolation and noise has been added to get low SNR image. The input image was Fourier transformed to get the spectrum image, binary image was resulted from the spectrum image and Fourier transformed again to obtain the correlation image, and the corresponding shift value was calculated with centre of mass calculation. The results show that the error does not change with different scene of input images, and the error RMS of 0.2 can be achieved.
出处
《计算机仿真》
CSCD
北大核心
2010年第3期215-219,共5页
Computer Simulation
关键词
联合变换相关器
像移测量
傅立叶变换
质量中心算法
Joint transform correlator
Image motion measurement
Fourier transform
Centre of mass calculation