Firstly, a new edge operator is proposed in defining the boundary of a 2-dimensional domain. Comparing this operator with others in some known denoising models shows that the boundary operator established in this paper is more efficient than the preceeding ones in detecting the edge of the domain. Secondly, based on the new edge operator and a new designed edge-stopping function, we improve the Alvarez equation and establish a new nonlinear isotropic diffusion image-denoising equation and we prove that the problems can be dealt with more efficiently with the newly established model than with others. Finally, the emulating numerical result obtained in this paper with the newly established model shows that the new model is indeed better than those gotten with the previous ones.