Diffusion coefficient-optimum SAR image speckle suppression n calculation method

Abstract

The invention discloses a diffusion coefficient-optimum SAR image speckle suppression calculation method. The method comprises the following steps of: calculating <epsilon>R (x, y) of a point column matrix of an SAR image; calculating a diffusion coefficient c (g); and carrying out speckle suppression on the SAR image. According to the method, the improved diffusion coefficient improves the disadvantage that the existing diffusion coefficients are inconsistent in the aspect of diffusion speed when being located in homogeneous areas; the diffusion coefficient can realize isotropic diffusion in the homogeneous areas, namely, c<eta>=c<zeta>=1, so that speckles can be removed preferably; and in heterogeneous areas, c<eta>=c<zeta>=1, the diffusion is stopped, so that a role of reverse smoothing can be played in the edges of the SAR image and then the edge information of the image is strengthened. The ratio of a diffusion speed in the <eta> direction to a diffusion speed in the >zeta> direction is zero, so that the diffusion coefficient satisfies a Charbonnier criterion. Experiments verify the effectiveness and superiority of the calculation method, and theoretically prove that the diffusion coefficient is stably iterated in the speckle suppression process.

Description

technical field [0001] The invention relates to an image processing technology, in particular to a calculation method for SAR image coherent speckle suppression. Background technique [0002] At present, many SAR image coherent speckle suppression methods based on anisotropic diffusion have been proposed, among which the classic anisotropic diffusion suppression methods mainly include: P-M anisotropic diffusion filter algorithm, SARD speckle suppression method, DPAD speckle suppression method method etc. [0003] The P-M anisotropic diffusion filter algorithm applies partial differential equations to the image. This method is proposed on the basis of additive noise. The traditional P-M diffusion is not suitable for the speckle suppression processing of SAR images directly because: one , P-M diffusion either cannot completely suppress the noise in the edge area, or it is easy to cause the loss of edge details; second, the gradient operator used by P-M diffusion is prone to n...

AI Technical Summary

Problems solved by technology

However, the Lee filter and Kuan filter coefficients generated by the local window estimation have no directionality and cannot match the local geometric features of the SAR image. They are static adaptive parameters.

Therefore, the diffusion filtering process of SRAD and DPAD will cause the unevenness of diffusion to be accumulated and amplified in multiple iterations, and eventually lead to serious block effects in local areas with high ICV values.

Therefore, the diffusion coefficient does not satisfy the Charbonnier criterion, and the diffusion speed in the two directions is different in the homogeneous region; although the diffusion coefficient is zero in the heterogeneous region, the diffusion speed is different

Images