Anisotropic Piecewise Gaussian Filter of Infrared Image
红外图像的各向异性分段高斯滤波

An anisotropic piecewise gaussian filter (PGF) has been putted forward. In this method, not only the information of scale and orientation has been applied, but also piecewise method has been used to preserve edge. In this article, a statistic model of noise has been used to estimate the variance of noise and based on this the standard scale of image has been decided. Then the Hermite transform method has been used to determine orientation adaptively. By calculating the axial ratio and scale of gaussian model, the effective neighbor filter can be converted into piecewise gaussian filter. The experimental result shows that this new method has better capability than other often used method in suppressing noise and preserving edge.