Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Anisotropic multi-directional total variation image denoising method and device

An anisotropic, multi-directional technology, applied in the field of image processing, can solve the problems of not considering the optimization of image oblique gradient information, not being able to better suppress image noise, and losing image detail information, so as to maintain edge feature information , to achieve the effect of optimized processing and good image quality

Active Publication Date: 2022-04-05
BEIJING JIAOTONG UNIV
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] At present, in the existing denoising methods based on total variation, some methods weight the pixels of the image before performing differential operations; some methods first smooth the total variation norm of the image, and then The total variation norm is iteratively solved, but this method will lose the detailed information of the image; some methods divide the image into local small blocks for local total variation filtering, but the computational complexity of this method is very high. time consuming to run
[0004] In short, the existing image denoising methods based on total variation do not take into account the optimization of the oblique gradient information of the image, and cannot better suppress the noise of the image.

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
  • Anisotropic multi-directional total variation image denoising method and device
  • Anisotropic multi-directional total variation image denoising method and device
  • Anisotropic multi-directional total variation image denoising method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0042] The embodiment of the present invention provides an image denoising method based on anisotropic multi-directional total variation, and the specific steps of the method are as follows:

[0043] Step 1: Obtain the original image to be denoised.

[0044] Step 2: Establish an anisotropic multi-directional total variation model of the original image, and perform a total variation regularization operation on the original image to be denoised. The above-mentioned multi-direction can include the horizontal, vertical and diagonal directions of the original image Perform a total variation difference operation.

[0045] The established anisotropic three-direction total variation model of the original image is as follows:

[0046]

[0047]Here u is the noise-free image to be optimized, and f is the original image polluted by noise. alpha h 、α v and alpha d are the coefficients of the total variation norm in the horizontal, vertical and diagonal directions respectively, and ...

Embodiment 2

[0085] This embodiment provides an anisotropic multi-directional total variation image denoising device, the structural diagram of which is as follows figure 2 As shown, the following modules are included:

[0086] Image acquiring unit 21, acquires the original image to be denoised;

[0087] An anisotropic multi-directional total variation model building unit 22, configured to perform a multi-directional total variation regularization operation on the original image, and establish an anisotropic multi-directional total variation model of the original image;

[0088] The optimization solving unit 23 is configured to optimize and solve the anisotropic multi-directional total variation model based on an iterative algorithm to obtain a denoised image of the original image.

[0089] Further, the anisotropic multi-directional total variation model building unit 22 is specifically used to perform total variation difference operations on the horizontal, vertical and diagonal directi...

Embodiment 3

[0098] The present invention will be further described by a group of examples below:

[0099] image 3 A schematic diagram of image comparison provided by an embodiment of the present invention. Select the Lena image and Barbara image as the experimental image, the image size is 64*64, the initial image is as follows image 3 -a and image 3 -e as shown in the image. Add Gaussian white noise to the initial image, such as image 3 -b and image 3 As shown in the -f image, the signal-to-noise ratio of the noisy image is 22.83dB.

[0100] Set up the model in the present invention, iterative optimization has the following calculation steps:

[0101] Iterative initial condition: when k=0, b 0 =04096×1 ,r 0 =0 4096×1 , u 0 =f 4096×1 , f 4096×1 =f 4096×1 ,0 represents an all-zero vector. Parameters in iterations μ=30, β=8, tol=10 -3 .

[0102] iteration loop:

[0103] In the first step, fix r and b, optimize u, and get u k+1 :

[0104]

[0105] In the second ste...

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 invention provides an anisotropic multi-directional total variation image denoising method and device. The steps of the method include establishing a multi-directional anisotropic total variation model, and performing regularization operations on the original image to be denoised (multi-directional differential operations such as horizontal difference, vertical difference, and diagonal line); The established anisotropic multi-directional total variation model is solved to obtain the denoised image. For the first time, the invention adds the full variation difference operation in the diagonal direction to the full variation model to denoise the image, and effectively optimizes the information such as the oblique direction of the image; it can effectively remove the influence of noise, and at the same time effectively maintain the image. The edge feature information provides good image quality for subsequent processing.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to an anisotropic multi-directional total variation image denoising method and device. Background technique [0002] The image is inevitably affected by noise during the acquisition process. The total variation (TV) image denoising method is currently an effective denoising method. This method regards the denoising process of the image as a piecewise constant model, The total variation model and iterative calculation are established to realize the denoising process of the image. However, the total variation image denoising method only optimizes the horizontal and vertical gradient information, and does not take into account the oblique information of the image, which makes the model lack the optimization of the oblique gradient information of the image. [0003] At present, in the existing denoising methods based on total variation, some methods weight the pixels of the im...

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
Patent Type & Authority Patents(China)
IPC IPC(8): G06T5/00
CPCG06T5/002G06T2207/20192
Inventor 申艳陈莹娄淑琴刘静郝晓莉侯亚丽陈后金闻映红张超余晶晶
Owner BEIJING JIAOTONG 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
Eureka Blog
Learn More
PatSnap group products