SAR image noise reduction method based on linear minimum mean square error estimation

A minimum mean square error, image noise reduction technology, applied in image enhancement, image data processing, calculation, etc., can solve the problem of loss of texture details

Inactive Publication Date: 2015-10-14
CHONGQING UNIV
View PDF2 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The object of the present invention is to propose a SAR image denoising method based on linear mini

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • SAR image noise reduction method based on linear minimum mean square error estimation
  • SAR image noise reduction method based on linear minimum mean square error estimation
  • SAR image noise reduction method based on linear minimum mean square error estimation

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0046] refer to figure 1 , the present invention is based on the SAR image denoising method of linear minimum mean square error estimation, and concrete steps comprise as follows:

[0047] Step 1. Clustering of similar image blocks

[0048] Firstly, the image block is extracted from the noisy image, and the image block set [y 1 ,y 2 ,...,y N ], then from [y 1 ,y 2 ,...,y N ], randomly select I image blocks as the initial class centers, and then for each image block in the similar block set, calculate the distance between them and the I class centers by formula (1), and attribute each image block to the class The class with the shortest distance from the center to itself, after all the image blocks are classified, the mean value of the set of image blocks of each class is used as the new class center, and the process of classifying the image blocks and updating the class center is repeated m times to obtain the final clustering result . After obtaining the initial set o...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The present invention discloses an SAR image noise reduction method based on linear minimum mean square error estimation. The method belongs to the technical field of digital image processing. The method is an SAR image noise reduction method that combines a nonlocal image similarity with a sparse representation. The method comprises: firstly, clustering similar images by a Kmeans clustering method; performing singular value decomposition on a similar block set to obtain a noisy singular value coefficient involving row and column correlation information; in order to enable the singular value coefficient after noise reduction to better approximate a reality coefficient, estimating the singular value coefficient by using a linear minimum mean square error criterion; and then reconstructing the estimated singular value coefficient to obtain an initial noise reduction image block, performing clustering noise reduction on noisy image blocks again in combination with an initial noise reduction result, and reconstructing the image blocks after noise reduction to obtain a final noise reduction image. According to the SAR image noise reduction method based on linear minimum mean square error estimation provided by the present invention, not only the noise reduction effect is obvious, but also the image texture details can be effectively preserved, and the final noise reduction image has a good visual effect; and the method can be applicable to SAR image noise reduction.

Description

technical field [0001] The invention belongs to the technical field of digital image processing, and in particular relates to a noise reduction method based on linear minimum mean square error estimation, which is used for SAR image noise reduction processing. Background technique [0002] Synthetic aperture radar has many advantages such as all-weather, all-weather, and side-looking imaging, so it is widely used in many fields, and it is playing an increasingly important role in military strikes, agricultural and forestry monitoring, and ocean development. However, the coherent speckle noise in the resulting SAR image not only affects the observation of the human eye, but also brings troubles to the subsequent image interpretation and target recognition. Therefore, noise reduction has become a core link in the pre-processing of SAR images. [0003] The traditional spatial filtering method not only has limited ability to suppress speckle noise, but also has serious loss of e...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06T5/00G06K9/62
Inventor 刘书君吴国庆张新征杨婷徐礼培
Owner CHONGQING UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap