Image edge detection scheme using wavelet transform

Abstract

The Wavelet Transform remained quite rapidly used technique today for analysing the signals. For image edge detection, wavelet transform provides facility to select the size of the image details that will be detected. Wavelets transform separates the lower frequencies and higher frequencies easily, which is prime important for edge detection. The wavelet scale sets the size of detected edges. For discrete wavelet transform, many signals are passed through wavelet filter for choice of the scale. For 2-D image, wavelet analysis is carried out in terms of horizontal and vertical function and edges are detected separately. In this paper the Daubachies wavelet transform has been used, where 2-D image is decomposed at three levels and at each level lower and higher frequencies have been separated. As Daubachies wavelet gives appropriate edge at three levels for black and white image along with some ghost edges. Some threshold has been used to cater these ghost edges. For edges detection in an image, a MATLAB code have been developed, which tests the two-dimensional scene. Two dimensional image objects have been investigated. The intention was to test 2D images by using wavelet transform of real objects.