Denoising Method of Remote Sensing Image Based on Scattering Matrix Features of Second Order Partial Differential Equation

Abstract

The invention discloses a second-order partial differential equation remote sensing image denoising method based on the characteristics of the scatter matrix, which belongs to the field of image processing. First, the control function of the diffusion term is established by introducing an operator with a constant 45N degree rotation modulus, so that the diffusion is The smoothing area is performed along the gradient direction and the vertical gradient direction, and only along the curvature direction in the edge area; secondly, the characteristic area of ​​the image is better resolved through the scatter matrix, and the fidelity item is constructed to improve the sharp corner The denoising effect of the area and smooth area effectively alleviates the over-smoothing effect of the MCM model at the sharp point of the target and the "blocky" effect produced by the PM model.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a second-order partial differential equation remote sensing image denoising method based on the characteristics of a scatter matrix, which has strong ability to maintain edges and sharp corners and high diffusion efficiency in smooth areas. Background technique [0002] Remote sensing image is a special type of image obtained by using the radiation characteristics of the observed object and the reflectivity of electromagnetic waves. It has become an important means of earth observation, involving many fields of national politics, economy, military and society and playing an important role. . However, because satellite remote sensing systems are often in complex electromagnetic wave environments, or due to factors such as electromagnetic interference between load components or periodic offsets of sensors, remote sensing images will be affected by noise to varying degrees during the...

AI Technical Summary

Problems solved by technology

However, the smoothing efficiency of this type of method for strong edge noise and isolated noise such as salt and pepper noise is not satisfactory, and many scholars have found that the PM method is "sick" and cannot guarantee the existence and uniqueness of the equation solution

For this reason, Catte et al. regularized the PM model to solve the uniqueness of the solution of the equation, but it is difficult to ensure the stability when the variance tends to 0

In order to better preserve the edge characteristics of the image, Alvarez et al. proposed the MCM (Mean Curvature Model) diffusion model based on the average curvature, so that the diffusion is always along the direction perpendicular to the gradient, but due to the diffusion speed and curvature of the model Proportional to , it will produce excessive smoothing in the sharp corner area of ​​the image

Images