CAJ | 학술논문

针对显微热图像，提出一种基于傅里叶变换的多分辨率位移估计方法。该方法首先对显微热图像进行小波变换，再将原始图像及小波变换后的各个近似分量生成一个基于灰度的金字塔序列图像，然后基于傅里叶变换位移估计算法分别对金字塔各层序列图像进行位移估计，最后将得到的各层序列图像位移量按照分层权重进行加权平均得到最后的图像序列位移量，实验结果表明：该方法有效地提高了显微热图像微位移估计精度，降低了位移估计的平均误差和方差。
침대현미열도상，제출일충기우부리협변환적다분변솔위이고계방법。해방법수선대현미열도상진행소파변환，재장원시도상급소파변환후적각개근사분량생성일개기우회도적금자탑서렬도상，연후기우부리협변환위이고계산법분별대금자탑각층서렬도상진행위이고계，최후장득도적각층서렬도상위이량안조분층권중진행가권평균득도최후적도상서렬위이량，실험결과표명：해방법유효지제고료현미열도상미위이고계정도，강저료위이고계적평균오차화방차。
According to micro thermal imaging, multiresolution displacement estimation based on Fourier transform was put forward. First micro thermal imaging was changed by wavelet transform, and a gray pyramid was built by the wavelet transform of image sequence. And the displacement components of each layer was estimated with Fourier transform. Then the final image sequence displacement was obtained by arithmetic layered weighted mean of all layers. The experimental results show that this method can effectively improve the micro thermal imaging displacement estimation accuracy, and reduce the average error and standard deviation mean.