CN110634112B - Method for enhancing noise-containing image under mine by double-domain decomposition - Google Patents

Method for enhancing noise-containing image under mine by double-domain decomposition Download PDF

Info

Publication number
CN110634112B
CN110634112B CN201910976955.3A CN201910976955A CN110634112B CN 110634112 B CN110634112 B CN 110634112B CN 201910976955 A CN201910976955 A CN 201910976955A CN 110634112 B CN110634112 B CN 110634112B
Authority
CN
China
Prior art keywords
image
noise
layer
detail
enhanced
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related
Application number
CN201910976955.3A
Other languages
Chinese (zh)
Other versions
CN110634112A (en
Inventor
田子建
王满利
张向阳
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China University of Mining and Technology Beijing CUMTB
Original Assignee
China University of Mining and Technology Beijing CUMTB
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by China University of Mining and Technology Beijing CUMTB filed Critical China University of Mining and Technology Beijing CUMTB
Priority to CN201910976955.3A priority Critical patent/CN110634112B/en
Publication of CN110634112A publication Critical patent/CN110634112A/en
Application granted granted Critical
Publication of CN110634112B publication Critical patent/CN110634112B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/90—Dynamic range modification of images or parts thereof
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/30—Noise filtering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20048—Transform domain processing
    • G06T2207/20064—Wavelet transform [DWT]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20212—Image combination
    • G06T2207/20221—Image fusion; Image merging

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Image Processing (AREA)

Abstract

The invention discloses a method for enhancing a dual-domain decomposed noise-containing image under a mine, which comprises the steps of firstly, using a spatial Gaussian filter to decompose the noise-containing image under the mine in a layered manner, and realizing decoupling of image contrast improvement and noise suppression; then, the maximum bright channel image of the basic layer obtained by layering is used as the illumination component estimation of the Retinex model, and the contrast of the basic layer is improved according to the Retinex enhancement principle; then, using a detail layer obtained by decomposing and layering non-downsampling shear wave transformation, performing hard threshold shrinkage on the decomposition coefficient in a shear wave transformation domain to realize detail layer noise suppression, obtaining a noise suppression detail layer by using the decomposition coefficient shrunk by non-downsampling shear wave inverse transformation, and realizing the enhancement of the noise suppression detail layer by using the same enhancement proportion of a basic layer; and finally, fusing the enhanced basic layer and the enhanced detail layer for noise suppression to obtain a fused enhanced image, and performing Gamma fine adjustment on the fused enhanced image to obtain a final enhanced image with improved noise suppression and contrast.

Description

Method for enhancing noise-containing image under mine by double-domain decomposition
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to a method for enhancing a noise-containing image under a mine by dual-domain decomposition.
Background
With the acceleration of development pace of intelligent mines, various image recognition technologies continuously extend to underground mines, but the underground mine imaging environment is poor, electromagnetic interference is high, the contrast of acquired images is low, and the acquired images are rich in noise, so that mine image enhancement and noise suppression become the primary problems to be solved in mine image recognition and application.
The current common algorithms for image enhancement are as follows: histogram equalization, histogram specification, image enhancement algorithms based on physical models, image enhancement algorithms based on partial differential equations and variations, and change domain image enhancement algorithms.
The histogram equalization enables the gray value of the image to occupy all possible gray levels, enables the equalized histogram to be distributed uniformly, improves the contrast of the image and improves the brightness of the image. The histogram normalization transforms the image histogram distribution into a desired distribution. The histogram specification is similar to the equalization algorithm and has the characteristics of simplicity and high efficiency, but the histogram conversion causes noise amplification, and therefore, the histogram specification cannot be sufficient for enhancing a noise image.
The enhancement algorithm based on the physical model is based on the image model, and the image disturbance component in the model is removed by utilizing the algorithm to realize the enhancement of the image, and the algorithm has definite physical significance. The Retinex model expresses an image as a product of a reflection component and an illumination component, wherein the reflection component represents the inherent color characteristic of an object, and the illumination component represents the brightness component of the environment. The image enhancement algorithm based on Retinex theory eliminates the illumination component in the model from the original image, solves the reflection component of the inherent color characteristic of the imaged object, and realizes image enhancement. However, the Retinex enhancement algorithm cannot identify the noise in the high frequency band, which is the same as the reflection component, so that it cannot be very good for enhancing the noise image.
The image enhancement algorithm based on partial differential equation and variation is to design different constraint conditions according to a certain image model, and optimize and solve the model to realize image enhancement. The enhancement effect of the algorithm is different along with the difference of the constraint conditions, and part of the enhancement algorithms have certain noise suppression capability, but the algorithm needs iterative solution and is complex, and the enhancement of the mine noisy image cannot be well performed.
The transform domain image enhancement algorithm is to decompose an image into decomposition coefficients in different frequencies and directions of a transform domain, strengthen the decomposition coefficients in different scales, reconstruct an enhanced image by utilizing the strengthened decomposition coefficients, and realize contrast improvement and detail feature highlighting. Compared with other types of enhancement algorithms, the transform domain image enhancement algorithm has unique advantages of separation and noise suppression, but the transform domain image enhancement algorithm has weak performance in the aspect of improving the contrast and brightness of an image, the underground mine imaging environment is poor, the image contrast is low, and the underground mine image enhancement algorithm has the strong capability of improving the contrast.
In view of the above, we have invented a method for enhancing a dual-domain decomposed mine underground noisy image. The method integrates the capability of improving the image contrast of the Retinex image enhancement method and the advantage of Non-Subsampled shear wave Transform (NSST) approximate to optimal sparse expression high-dimensional function separation noise, decouples the image contrast improvement and noise suppression under the mine, and achieves the purpose of enhancing the noise-containing image under the mine.
Disclosure of Invention
In order to improve the observability of noise-containing images under a mine, the invention utilizes the airspace mean filter to layer noise images, realizes the decoupling of a basic layer for determining the image contrast and a detail layer containing image texture details and noise, eliminates the interference of the noise when the image contrast is improved, eliminates the influence of noise inhibition on the improvement of the contrast, and realizes the contrast improvement and noise inhibition of the noise images under the mine by decoupling.
In order to achieve the purpose, the invention adopts the technical scheme that: a method for enhancing a noise-containing image under a mine by dual-domain decomposition comprises the following steps:
the method comprises the following steps: spatial decomposition of the image: three color channels f of noise-containing images under mine by utilizing airspace Gaussian filtercC belongs to { r, g, b }, layering, and the output of the Gaussian filter is a basic layer
Figure BDA0002233954360000021
The difference image of the original noise image and the basic layer is a detail layer
Figure BDA0002233954360000022
The base layer determines the overall contrast of the image, and the detail layer comprises texture details and noise of the original image;
step two: reinforcing a base layer: computing a base layer according to a Retienex model
Figure BDA0002233954360000023
I.e. calculating the maximum value of the three color channels, and taking the maximum value of the bright channel as the RetienIllumination component L of ex modelbase，
Figure BDA0002233954360000024
Dividing by the illumination component to obtain an enhanced base layer image
Figure BDA0002233954360000025
Step three: non-downsampling shear wave transformation noise reduction of detail layers: slicing layers using non-subsampled shear wave transforms
Figure BDA0002233954360000026
c belongs to { r, g, b } decomposition, and the obtained decomposition coefficient matrix
Figure BDA0002233954360000027
c∈{r,g,b},j＝1,2,…,J,s＝1,2,…,SjJ is the maximum resolution scale number, SjFor the maximum number of directions corresponding to the j-scale, the threshold shrinks
Figure BDA0002233954360000028
Obtaining a coefficient matrix for noise suppression
Figure BDA0002233954360000029
To pair
Figure BDA00022339543600000210
Reconstructing noise suppressed detail layers by applying non-subsampled inverse shear wave transform
Figure BDA00022339543600000211
Is divided by LbaseDeriving a noise suppression enhanced detail layer
Figure BDA00022339543600000212
Step four: layer image fusion: enhancing noise suppression by detail layer
Figure BDA00022339543600000213
Multiplied by a detail enhancement factor λ, λ ≧ 1 plus
Figure BDA00022339543600000214
Obtaining a fused enhanced image
Figure BDA00022339543600000215
Fine tuning using Gamma transformation
Figure BDA00022339543600000216
Obtaining a final enhanced image
Figure BDA00022339543600000217
Further, according to the method for enhancing the noise-containing image under the mine with the dual-domain decomposition, in the step one, the gaussian kernel function of the gaussian filter is
Figure BDA00022339543600000218
Where σ is the variance of the gaussian kernel function.
The r, g, b channel base layer image is obtained by convolving an input mine underground noise image with a Gaussian kernel function, and then the detail layer image is obtained by subtracting the input image from the base layer:
Figure BDA0002233954360000031
wherein denotes a convolution operation fc、
Figure BDA0002233954360000032
And c belongs to the { r, g, b } and represents three color channels of the color image.
Further, according to the method for enhancing the noise-containing image under the mine with the dual-domain decomposition, in the second step, the base layer
Figure BDA0002233954360000033
c is the maximum bright channel of r, g, b
Figure BDA0002233954360000034
As an estimate of the illumination component of the Retinex model, where ε is the prevention of LbaseA small positive number of zero, the enhanced image of the base layer
Figure BDA0002233954360000035
From the Retinex model:
Figure BDA0002233954360000036
furthermore, according to the method for enhancing the noise-containing image under the mine with the dual-domain decomposition, in the third step, for convenience of expression, a symbol T is definedNSST() Representing NSST transforms, the detail layer decomposition coefficient sets
Figure BDA0002233954360000037
Figure BDA0002233954360000038
From detail layer images
Figure BDA0002233954360000039
By NSST decomposition, the following results are obtained:
Figure BDA00022339543600000310
wherein J is the maximum resolution scale number, SjThe maximum number of directions corresponding to the j-scale.
Decomposition coefficients according to detail layer
Figure BDA00022339543600000311
Noise-suppressed detail layer decomposition coefficients
Figure BDA00022339543600000312
By hard threshold shrinkage we derive:
Figure BDA00022339543600000313
in the formula, kjRepresenting the threshold coefficient, σ, corresponding to the scale jcExpressing the standard deviation of noise, estimated by wavelet noise empirical estimation formula, using Sym4 wavelet pair detail layer image of Symlets wavelet system
Figure BDA00022339543600000314
Performing a decomposition using diagonal detail wavelet coefficients
Figure BDA00022339543600000315
Median estimated noise standard deviation σc:
Figure BDA00022339543600000316
In the formula, | · | is an absolute value operator, and mean (-) represents a median operation.
Figure BDA00022339543600000317
The variance estimation of the decomposition coefficient matrix corresponding to different scales j and the directional filter s obtained by combining a Monte Carlo method (Monte Carlo method) with a wavelet noise empirical formula is shown.
Figure BDA00022339543600000318
And (3) calculating: first, a size and is generated by a pseudo-random algorithm
Figure BDA00022339543600000319
Noise image Noise of the same unit standard deviation1(ii) a Then, using the atrous wavelet to decompose with NSST
Figure BDA0002233954360000041
The number of scales and the number of directional filters decompose Noise1(ii) a Finally, estimate Noise using wavelet Noise empirical formula1Noise variance of each decomposition coefficient
Figure BDA0002233954360000042
According to
Figure BDA0002233954360000043
Reconstructed from NSST inverse transform
Figure BDA0002233954360000044
Noise suppression enhancement detail layer
Figure BDA0002233954360000045
By
Figure BDA0002233954360000046
Is divided by LbaseTo obtain:
Figure BDA0002233954360000047
in the formula (I), the compound is shown in the specification,
Figure BDA0002233954360000048
representing the NSST inverse transform.
Further, according to the method for enhancing the noise-containing image under the mine with the double-domain decomposition, in the fourth step, the image is enhanced by layer fusion
Figure BDA0002233954360000049
By
Figure BDA00022339543600000410
And
Figure BDA00022339543600000411
according to a fusion formula
Figure BDA00022339543600000412
Lambda is more than or equal to 1, and finally the image is enhanced
Figure BDA00022339543600000413
Corrected by Gamma
Figure BDA00022339543600000414
To obtain:
Figure BDA00022339543600000415
in the formula (I), the compound is shown in the specification,
Figure BDA00022339543600000416
and
Figure BDA00022339543600000417
to represent
Figure BDA00022339543600000418
Minimum and maximum values of.
The invention has the beneficial effects that: aiming at the contradiction between contrast improvement and noise suppression when a noise image under a mine is enhanced, a spatial domain Gaussian filter is used for decomposing an image in a layered mode, and decoupling of the image contrast improvement and the noise suppression is achieved; enhancing the base layer image according to the Retinex model, and improving the contrast of the base layer image; the advantage of NSST approximate optimal sparse expression high-dimensional function is used, the detail layer image is sparsely expressed, and the decomposition coefficient of NSST is shrunk through a threshold value, so that the detail layer noise suppression is realized; the images of the base layer and the detail layer after the fusion processing achieve the purposes of improving the contrast of the whole image and suppressing noise.
Drawings
FIG. 1 is an enhancement framework for dual domain decomposition of a noise image in a mine.
Detailed Description
The invention is described in detail below with reference to the drawings and the examples.
A method for enhancing a mine underground noise-containing image by dual-domain decomposition comprises the following steps:
the first step is as follows: input image r, g, b channel layered decomposition
As shown in FIG. 1, the input image is hierarchically decomposed in three channels of r, g and b by using a Gaussian filter, and the high channelsThe gaussian filter is a spatial smoothing filter, the template size of the gaussian kernel function of the gaussian filter is 15 × 15, and the variance is 3. In a broken line frame of 'layered decomposition of r, g and b channels of images', r, g and b channel images of a base layer
Figure BDA0002233954360000051
From the input image fcC is the { r, g, b } and Gaussian kernel function
Figure BDA0002233954360000052
Convolution is carried out to obtain the images of the r, g and b channels of the detail layer from fcC is equal to { r, g, b } minus
Figure BDA0002233954360000053
To obtain that:
Figure BDA0002233954360000054
the second step is that: base layer r, g, b channel enhancement
As shown in fig. 1, the channel enhancement of the base layer r, g, b is obtained by layered decomposition according to the gaussian filter in the dashed box
Figure BDA0002233954360000055
Get
Figure BDA0002233954360000056
As the illumination component of Retinex, the maximum of the r, g, b channels of (1):
Figure BDA0002233954360000057
epsilon is a small positive number with unequal zeros, the effect being to prevent LbaseIs zero. According to the Retinex model:
Figure BDA0002233954360000058
it can be seen that the enhanced image of the base layer
Figure BDA0002233954360000059
Can be composed of
Figure BDA00022339543600000510
Is divided by LbaseObtaining, namely:
Figure BDA00022339543600000511
the third step: fine layer r, g, b channel noise reduction
As shown in FIG. 1, in the dotted line frame of "noise reduction of r, g, b channels of detail layer", the detail layer image obtained by hierarchical decomposition
Figure BDA00022339543600000512
NSST adopts 4 scales, and the corresponding directional filter number is 8, 16 and 16 to decompose the r, g and b channel images of the detail layers to obtain
Figure BDA00022339543600000513
Set of decomposition coefficients for c ∈ { r, g, b }
Figure BDA00022339543600000514
Hard threshold shrinkage using threshold coefficients of 0, 3.1 and 4
Figure BDA00022339543600000515
Set of decomposition coefficients resulting in noise suppression
Figure BDA00022339543600000516
Figure BDA00022339543600000517
Wherein the noise standard deviation sigmacEstimated by wavelet noise empirical estimation formula
Figure BDA00022339543600000518
c is in the range of r, g, b, and the variance of the decomposition coefficient matrix of the detail layer
Figure BDA00022339543600000519
Estimated by the Monte Carlo method,
Figure BDA00022339543600000520
and (3) calculating: first, a size and is generated by a pseudo-random algorithm
Figure BDA00022339543600000521
Noise image Noise of the same unit standard deviation1(ii) a Then, using the atrous wavelet to decompose with NSST
Figure BDA00022339543600000522
The number of scales and the number of directional filters decompose Noise1(ii) a Finally, estimate Noise using wavelet Noise empirical formula1Noise variance of each decomposition coefficient
Figure BDA00022339543600000523
Then, for
Figure BDA00022339543600000524
Figure BDA00022339543600000525
Detail layer with noise suppression by performing inverse NSST transform
Figure BDA00022339543600000526
Finally, according to the enhancement proportion of the basic layer, the detail layer for enhancing noise suppression
Figure BDA00022339543600000527
Namely, it is
Figure BDA00022339543600000528
Is divided by LbaseDeriving a noise-suppressing enhancement detail layer
Figure BDA00022339543600000529
Namely:
Figure BDA0002233954360000061
the fourth step:
as shown in FIG. 1, in the dashed box of "layered r, g, b image fusion", an enhanced base layer is used
Figure BDA0002233954360000062
And a noise enhancement detail layer
Figure BDA0002233954360000063
According to the fusion rule of the layer images:
Figure BDA0002233954360000064
obtaining a layer fusion enhanced image with the detail enhancement coefficient lambda being 1.4
Figure BDA0002233954360000065
To pair
Figure BDA0002233954360000066
Implementing Gamma conversion to obtain the final enhanced image
Figure BDA0002233954360000067
Namely:
Figure BDA0002233954360000068

Claims (5)

1. a method for enhancing a noise-containing image under a mine by dual-domain decomposition is characterized by comprising the following steps:
the method comprises the following steps: spatial decomposition of the image: three color channels f of noise-containing images under mine by utilizing airspace Gaussian filtercC is in the hierarchy of r, g, b, and the output of the Gaussian filter is the basic layer
Figure FDA0003385420180000011
The difference image of the original noise image and the basic layer is a detail layer
Figure FDA0003385420180000012
FoundationThe layer determines the overall contrast of the image, and the detail layer comprises the texture detail and the noise of the original image;
step two: reinforcing a base layer: computing a base layer according to a Retienex model
Figure FDA0003385420180000013
The maximum bright channel value is used as the illumination component of the Reinex model, namely the maximum value of the three color channels is calculated
Figure FDA0003385420180000014
Dividing by the illumination component to obtain an enhanced base layer image
Figure FDA0003385420180000015
Step three: non-downsampling shear wave transformation noise reduction of detail layers: slicing layers using non-subsampled shear wave transforms
Figure FDA0003385420180000016
Decomposing to obtain a decomposition coefficient matrix
Figure FDA0003385420180000017
J is the maximum resolution scale number, SjFor the maximum number of directions corresponding to the j-scale, the threshold shrinks
Figure FDA0003385420180000018
Obtaining a coefficient matrix for noise suppression
Figure FDA0003385420180000019
To pair
Figure FDA00033854201800000110
Reconstructing noise suppressed detail layers by applying non-subsampled inverse shear wave transform
Figure FDA00033854201800000111
Figure FDA00033854201800000112
Is divided by LbaseDeriving a noise suppression enhanced detail layer
Figure FDA00033854201800000113
Step four: layer image fusion: enhancing noise suppression by detail layer
Figure FDA00033854201800000114
Multiplied by a detail enhancement factor λ, λ ≧ 1 plus
Figure FDA00033854201800000115
Obtaining a fused enhanced image
Figure FDA00033854201800000116
Fine tuning using Gamma transformation
Figure FDA00033854201800000117
Obtaining a final enhanced image
Figure FDA00033854201800000118
2. The method for enhancing noise-containing images in mine wells through dual-domain decomposition according to claim 1, wherein in the step one, the Gaussian kernel function of the Gaussian filter is
Figure FDA00033854201800000119
Where σ is the variance of the gaussian kernel function. The r, g, b channel base layer image is obtained by convolving an input mine underground noise image with a Gaussian kernel function, and then the detail layer image is obtained by subtracting the input image from the base layer:
Figure FDA00033854201800000120
wherein denotes a convolution operation fc、
Figure FDA00033854201800000121
And c belongs to the { r, g, b } and represents three color channels of the color image.
3. The method for enhancing the noise-containing image under the mine with the dual-domain decomposition as claimed in claim 1 or 2, wherein in the second step, the base layer
Figure FDA00033854201800000122
Maximum bright channel of
Figure FDA00033854201800000123
As an estimate of the Retinex mode illumination component, where ε is the prevention of LbaseSmall positive, zero, enhancement image of base layer
Figure FDA00033854201800000124
Derived from Retinex model
Figure FDA0003385420180000021
4. The method for enhancing the noise-containing image under the mine with the double-domain decomposition as claimed in claim 3, wherein in the third step, the detail layer is decomposed by the non-downsampling shear wave transformation to obtain the detail layer decomposition coefficient set
Figure FDA0003385420180000022
Figure FDA0003385420180000023
The hard threshold of the decomposition coefficient is shrunk,set of decomposition coefficients resulting in noise suppression
Figure FDA0003385420180000024
Figure FDA0003385420180000025
Reconstructing a fused enhanced image from a set of non-subsampled shear wave inverse transform noise suppressed decomposition coefficients
Figure FDA0003385420180000026
Fusing enhanced images
Figure FDA0003385420180000027
Divided by the basic layer
Figure FDA0003385420180000028
Maximum bright channel L ofbaseA final reinforcing detail layer is obtained.
5. The method for enhancing noise-containing images under mine of claim 4, wherein in the fourth step, the image is enhanced by layer fusion
Figure FDA0003385420180000029
By
Figure FDA00033854201800000210
And
Figure FDA00033854201800000211
according to a fusion formula
Figure FDA00033854201800000212
Lambda is more than or equal to 1, and finally the image is enhanced
Figure FDA00033854201800000213
Corrected by Gamma
Figure FDA00033854201800000214
To obtain:
Figure FDA00033854201800000215
in the formula (I), the compound is shown in the specification,
Figure FDA00033854201800000216
and
Figure FDA00033854201800000217
to represent
Figure FDA00033854201800000218
Minimum and maximum values of.
CN201910976955.3A 2019-10-15 2019-10-15 Method for enhancing noise-containing image under mine by double-domain decomposition Expired - Fee Related CN110634112B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910976955.3A CN110634112B (en) 2019-10-15 2019-10-15 Method for enhancing noise-containing image under mine by double-domain decomposition

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910976955.3A CN110634112B (en) 2019-10-15 2019-10-15 Method for enhancing noise-containing image under mine by double-domain decomposition

Publications (2)

Publication Number Publication Date
CN110634112A CN110634112A (en) 2019-12-31
CN110634112B true CN110634112B (en) 2022-02-22

Family

ID=68975120

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910976955.3A Expired - Fee Related CN110634112B (en) 2019-10-15 2019-10-15 Method for enhancing noise-containing image under mine by double-domain decomposition

Country Status (1)

Country Link
CN (1) CN110634112B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113674158A (en) * 2020-05-13 2021-11-19 浙江宇视科技有限公司 Image processing method, device, equipment and storage medium
CN111898567B (en) * 2020-08-04 2023-11-07 昆山九毫米电子科技有限公司 Shot shock wave signal noise suppression method based on empirical wavelet transformation
CN117372301A (en) * 2023-09-27 2024-01-09 深圳市梦作坊科技有限公司 Game image processing method based on multimode coupling technology

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104346776A (en) * 2013-08-02 2015-02-11 杭州海康威视数字技术股份有限公司 Retinex-theory-based nonlinear image enhancement method and system
CN105139367A (en) * 2015-07-27 2015-12-09 中国科学院光电技术研究所 Visible light polarization image fusion method based on non-subsampled shear wave
CN107798665A (en) * 2017-11-07 2018-03-13 天津大学 Underwater picture Enhancement Method based on structural texture layering
CN107862666A (en) * 2017-11-22 2018-03-30 新疆大学 Mixing Enhancement Methods about Satellite Images based on NSST domains
CN109961411A (en) * 2019-03-12 2019-07-02 浙江工业大学之江学院 Non-lower sampling shears wave conversion medicine CT image denoising method

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105894462B (en) * 2016-03-21 2019-01-11 辽宁师范大学 Remote sensing images denoising method based on shearing wave zone Hidden markov tree model
CN109146814B (en) * 2018-08-20 2021-02-23 Oppo广东移动通信有限公司 Image processing method, image processing device, storage medium and electronic equipment
CN109919861A (en) * 2019-01-29 2019-06-21 浙江数链科技有限公司 Infrared image enhancing method, device, computer equipment and storage medium
CN110175964B (en) * 2019-05-30 2022-09-30 大连海事大学 Retinex image enhancement method based on Laplacian pyramid

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104346776A (en) * 2013-08-02 2015-02-11 杭州海康威视数字技术股份有限公司 Retinex-theory-based nonlinear image enhancement method and system
CN105139367A (en) * 2015-07-27 2015-12-09 中国科学院光电技术研究所 Visible light polarization image fusion method based on non-subsampled shear wave
CN107798665A (en) * 2017-11-07 2018-03-13 天津大学 Underwater picture Enhancement Method based on structural texture layering
CN107862666A (en) * 2017-11-22 2018-03-30 新疆大学 Mixing Enhancement Methods about Satellite Images based on NSST domains
CN109961411A (en) * 2019-03-12 2019-07-02 浙江工业大学之江学院 Non-lower sampling shears wave conversion medicine CT image denoising method

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
Remote sensing image enhancement based on the combination of nonsubsampled shearlet transform and guided filtering;Duliang Lv;《Optical Engineering》;20161031;全文 *
基于双域分解的矿井下图像增强算法;田子建 等;《光子学报》;20190531;第48卷(第5期);全文 *

Also Published As

Publication number Publication date
CN110634112A (en) 2019-12-31

Similar Documents

Publication Publication Date Title
Majumdar Blind denoising autoencoder
CN110634112B (en) Method for enhancing noise-containing image under mine by double-domain decomposition
Diwakar et al. CT image denoising using NLM and correlation‐based wavelet packet thresholding
CN101944230B (en) Multi-scale-based natural image non-local mean noise reduction method
Jardim et al. Savitzky–Golay filtering as image noise reduction with sharp color reset
Kollem et al. Denoising and segmentation of MR images using fourth order non‐linear adaptive PDE and new convergent clustering
CN104820969A (en) Real-time blind image restoration method
Grigoryan et al. Optimal Wiener and homomorphic filtration
Shahdoosti et al. Combined ripplet and total variation image denoising methods using twin support vector machines
Diwakar et al. CT image denoising based on complex wavelet transform using local adaptive thresholding and Bilateral filtering
CN114418886A (en) Robustness denoising method based on deep convolution self-encoder
Krishnan et al. An efficient operator splitting method for noise removal in images
CN105096274B (en) Infrared image noise-reduction method based on non-down sampling contourlet domain mixing statistical model
CN108649926B (en) DAS data denoising method based on wavelet basis tensor sparse representation
Zhang et al. Image denoising based on the wavelet semi-soft threshold and total variation
Qingrong et al. Contrast enhancement of medical images using fuzzy set theory and nonsubsampled shearlet transform
Wali et al. Fast and adaptive boosting techniques for variational based image restoration
Li et al. A novel remote sensing image enhancement method using unsharp masking in nsst domain
Deivalakshmi et al. Balanced GHM Mutiwavelet Transform Based Contrast Enhancement Technique for Dark Images Using Dynamic Stochastic Resonance.
Iqbal et al. Reduction of speckle noise in medical images using stationary wavelet transform and fuzzy logic
Rela et al. Efficient image enhancement techniques applied on medical imaging-A state-of-the art survey
Liu et al. The validity analysis of the non-local mean filter and a derived novel denoising method
Tseng et al. Image restoration using hybrid features improvement on morphological component analysis
Mahesh et al. An Effective Image Denoising Using Adaptive Thresholding In Wavelet Domain
Kumar et al. Edge preservation based CT image denoising using wavelet and curvelet transforms

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20220222

CF01 Termination of patent right due to non-payment of annual fee