SAR interferogram wavelet noise reduction processing method and processing device thereof

Abstract

The invention provides an SAR interferogram wavelet noise reduction processing method comprising the following steps: determining signal points and noise points in an SAR interferogram, constructing a noise point identification matrix, and only carrying out filtering noise reduction processing on the noise points; the method comprises the following steps of: extracting interference pattern data containing noise, decomposing the interference pattern data into a real part and an imaginary part, respectively performing wavelet decomposition on the real part and the imaginary part by utilizing a Biorthogonal wavelet basis, and enabling the number of layers of wavelet decomposition to be optimal by adopting a whitening test mode; carrying out noise reduction processing on a wavelet coefficient by utilizing an improved wavelet threshold function and a hierarchical threshold value obtained by a Birge-Massart penalty function in a self-adaptive manner; and wavelet reconstruction is carried out on the real part and the imaginary part after filtering and noise reduction, an interferogram after filtering and noise reduction is obtained, and a final noise reduction interferogram is obtained by combining signal points in the SAR interferogram. The invention further provides an SAR interferogram wavelet noise reduction processing device. Noise can be effectively reduced, and edge detail information can be kept.

Description

technical field [0001] The invention relates to the field of image processing. More specifically, the present invention relates to a SAR interferogram wavelet noise reduction processing method and a processing device thereof. Background technique [0002] Due to the difference in the distance between the scatterer and the sensor, the coherence of the echo signal is affected, and the SAR (Synthetic Aperture Radar, Synthetic Aperture Radar) imaging produces coherent speckle noise, which reduces the quality of the SAR image to a certain extent and can lead to The obtained features are distorted. Therefore, filtering and denoising SAR images can improve the interpretation ability of SAR images and obtain more real and reliable phase information. [0003] At present, there are two main types of SAR noise reduction technologies: one is multi-view processing technology before imaging, and the other is filtering processing technology after imaging. The multi-view processing techn...

AI Technical Summary

Problems solved by technology

Saravani et al. proposed a new wavelet transform iterative algorithm, which denoises SAR images by aggregating stationary wavelet transform, bilateral filtering, Bayesian estimation and anisotropic diffusion filtering, which effectively reduces speckle noise and preserves image quality. The edge detail information; but the iterative algorithm has too much calculation, which may cause low efficiency

Zhang Yuye and others used stationary wavelet to decompose the signal, and used enhanced Lee filtering and adaptive threshold filtering methods to process different signals separately, and finally achieved the purpose of noise reduction; but the combination of the two filtering processes is relatively complicated, and there is no wavelet filtering. Refine and improve

Yan et al. obtained the wavelet packet coefficients by using fast non-local mean filtering, and performed wavelet packet transformation on the real and imaginary parts of the SAR image, which avoided the influence of phase jump on subsequent filtering and effectively improved the quality of noise reduction; but increased The complexity of the algorithm

[0006] In summary, SAR image noise reduction has been a long-term research hotspot in the field of radar remote sensing. In the past, it was difficult for traditional filtering methods to take into account effective noise reduction while maintaining edge detail information and features.

In recent years, wavelet noise reduction has been widely used as a new filtering method. Although many scholars have used different research ideas to improve the effect of wavelet noise reduction, most algorithms often have the following limitations: (1) The algorithm is relatively complex, resulting in a large amount of calculation. (2) The algorithm introduces other variables, which increases the unreliability; (3) The details of decomposition, noise reduction and reconstruction in the wavelet denoising method are not summarized and studied enough

Images