高级检索

利用Gabor滤波的相位图像进行光流估计

Estimation of Optical Flow for Phase Image Using Gabor Filter Banks

  • 摘要: 提出一种联合Gabor滤波器和加权最小二乘法估计序列图像光流的方法,用Gabor滤波器的滤波结果作为新的图像特征来计算空间和时间相位梯度.为了获得较为精确的结果,用加权最小二乘法来估计序列图像的光流.实验结果表明,文中算法在光流的计算精度上要优于Lucas,Fleet等的方法.

     

    Abstract: In the paper, a new algorithm for estimating the optical flow of image sequences is presented, by combining Gabor filter banks with weighted least-squares. The results of Gabor filter banks are used as the new features of image for the computation of the spatial-temporal phase gradient. In order to obtain more accurate results, the optical flow of sequence image is estimated by the method of weighted least-squares. The results of the experiments indicate that more accurate results of optical estimation can be obtained compared with the methods suggested by Lucas, Fleet, et al.

     

/

返回文章
返回