This paper shows that in edge detection the regularization factor a is a be
tter scale parameter than the standard deviation (sigma) of the Gaussian pr
e-filter. The a scale space, which exhibits the evolutionary behavior of an
edge in various scales, is the basis for the design of a multiscale edge d
etector (MRCBS). In MRCBS, the scale is determined adaptively according to
the local noise level; the thresholds which control the amount of edge deta
ils are adjusted according to the scale; and the anisotropic diffusion is a
pplied in the finest scale to further suppress noise.