Abstract: Aiming at the problem of the isotropic characteristics existed while dealing with image edges through traditional Gaussian-Laplace (LOG) operator, an anisotropic LOG algorithm is proposed to adaptively process image edge of various angles during image enhancement. Firstly, the scale parameters are introduced into the function by taking different variances in different directions of the LOG function. Then, combining with the geometry, the angle parameters are introduced into the operator with the scale parameters. Finally, the scale parameters are determined according to the gradient values of each pixel. The value of the angle parameter gives the anisotropic LOG operator. The simulation results show that the anisotropic LOG operator is superior to the isotropic LOG operator in terms of peak signal-to-noise ratio, structural similarity and mean square error. Thus the anisotropic LOG operator has more suppression effect on noise while retaining more image edge details.