用离散傅立叶-切比雪夫正交矩描述图像

Image Analysis by Fourier-Chebyshev Moments

提出了一种基于离散Fourier正交函数系与离散Chebyshev正交多项式的图像正交矩。分析了连续正交矩与离散正交矩的原理与重构方法,提出Fourier Chebyshev正交矩的构造原理与图像重构方法。对Legendre矩与Zernike矩两种经典的图像矩与离散Fourier Chebyshev正交矩做了对比实验。实验结果表明,在复杂图像的处理中,该矩具有比传统连续函数矩更好的特征表达能力。

A new set of orthogonal moment functions were introduced based on the discrete Fourier functions and the discrete Chebyshev polynomials. The Fourier-Chebyshev moments can be effectively used as image features in the analysis of two-dimensional images. The paper constructs the moments in the discrete orthogonal domain of the image coordinate space. The property makes the images be reconstructed by no more than N x N moments. The experimental results indicate that the Fourier-Chebyshev moments are superior to the conventional moments such as Legendre moments and Zernike moments in the analysis of complex images.