Application Of Wavelet Transform In Edge Detection

Wavelets have became an important research field in signal processing due to its unique advantage in non-stationary signal processing. With the development of wavelets theory, wavelets have been given more and more attention in image edge detection for its time-frequency localization and multi-scale property.This paper first summarizes the traditional edge detect operation, then intro-duce the Modulus Maxima Algorithm based on Discrete Wavelet Transform(DWT) and Multi-scale Edge Detection. In fact, Much image pre-processing such as edge detection and text analysis requires nearly shift invariance at least, but DWT is shift sensitive, so this paper studies the shift invariance property of Stationary Wavelet Transform(SWT), and do a series of experiment about edge detection based on SWT,its results are better than experiments based on DWT. Otherwise, DWT has poor direction analysis ability (only can provide three spatial orientations), while Dual-tree Complex Wavelet Transform(DT - CWT) can provide six spatial orien-tations details; including±15°,±45°.±75°. This paper puts forward an edge detection method, which combines DWT and DT - CWT, the experiment results presents its good performance.