An Image Edge Detection Algorithm Based on Wavelet Transform and Mathematical Morphology

Abstract

Edge detection is an important step in image processing and analysis. Traditional edge detection operators are sensitive to noises. Using wavelet transform to image edge detection can restrain the noise very good, there is a phenomenon of discontinuous at the edge of the detected image. And the one based on mathematical morphology can extract relatively continuous and smooth edges, but it often extracts the thicker edge. This paper proposes an algorithm which based on wavelet transform and mathematical morphology. It chooses an improved multi-structural anti-noise morphology edge detection operator to perform low-frequency edge extraction, and uses wavelet modulus maxima edge extraction on the high frequency components, the results of the above operations are fused into the final outcome. The experimental results show that, compared with the traditional edge detection operators, the proposed algorithm can effectively suppress noises and improve detection accuracy as well as positioning accuracy.

Heilongjiang Provincial Department of Education Foundation (Grant No.: 12521057) funded project.