基于加窗Hilbert变换的高精度边缘检测

常规Hilbert变换对噪音很敏感,目前主要用于三瞬参数的计算。随着不断的研究,现已发展了多种广义Hilbert变换法,这些方法抗噪能力强、对边缘敏感性更高。本文介绍了频率域加窗Hilbert变换,提出了一套新的基于时间域加窗的Hilbert变换边缘检测算法。实例分析结果表明,该方法边缘检测效果与频率域加窗Hilbert变换相当,在希氏因子长度较短时其计算效率明显高于频率域的。

High accuracy edge detection on window-based Hilbert Transform

Hilbert Transform is sensitive to noise, so it is extensively applied to calculate instantaneous characters now. With continuous the research and development, now many generalized Hilbert Transform has been proposed. It has better robustness than the traditional Hilbert Transform and it is more sensitive to edge. In this paper, window-based Hilbert Transform in frequency domain is introduced and a new algorithm of edge detection window-based Hilbert Transform is proposed. Not only has it the same capacity in edge detection as Hilbert Transform with window in frequency domain, but also has it a higher calculation speed.