摘要
研究了运动模糊图像在空间域和频域上的数学模型,理论分析指出物体作高频谐波振动时运动模糊成像的频谱上会由于Bessel函数的零点的存在而形成系列暗条纹。提出利用CCD摄取一幅振动模糊图像,经二维傅里叶变换,并采用图像处理技术检测频谱上直线条纹的位置,即可实现高频振动振幅的图像测量,测量结果为亚像素级的精度。
Mathematical modelal are researched in space domain and frequency domain for motion-blurred image. Series of dark stripes can be found on spectrums of motion-blurred images of a vibrating object with high frequency, which is due to zero value of Bessel function. Adding 2-D Fourier transform on a vibration-blurred image recorded by CCD, linear stripes are inspected and located in spectrum images and realize the image measurement of vibration amplitudes. Measurement results can reach sub-pixel precision.
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
2004年第4期85-88,共4页
Journal of Mechanical Engineering
基金
国家自然科学基金(50390063)
博士点基金(20020248048)
