CN108876741A - A kind of image enchancing method under the conditions of complex illumination - Google Patents
A kind of image enchancing method under the conditions of complex illumination Download PDFInfo
- Publication number
- CN108876741A CN108876741A CN201810647661.1A CN201810647661A CN108876741A CN 108876741 A CN108876741 A CN 108876741A CN 201810647661 A CN201810647661 A CN 201810647661A CN 108876741 A CN108876741 A CN 108876741A
- Authority
- CN
- China
- Prior art keywords
- image
- fuzzy
- high frequency
- frequency coefficient
- coefficient
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000005286 illumination Methods 0.000 title claims abstract description 26
- 238000012545 processing Methods 0.000 claims abstract description 45
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 24
- 230000000694 effects Effects 0.000 claims abstract description 9
- 238000000605 extraction Methods 0.000 claims abstract description 9
- 239000003245 coal Substances 0.000 claims abstract description 7
- 238000001914 filtration Methods 0.000 claims abstract description 3
- 238000013461 design Methods 0.000 claims description 12
- 238000010276 construction Methods 0.000 claims description 7
- 230000002708 enhancing effect Effects 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000003384 imaging method Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 230000001105 regulatory effect Effects 0.000 claims description 3
- 238000006467 substitution reaction Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 101100001678 Emericella variicolor andM gene Proteins 0.000 claims description 2
- 239000000284 extract Substances 0.000 claims description 2
- 230000005540 biological transmission Effects 0.000 abstract description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000004321 preservation Methods 0.000 description 2
- 230000003014 reinforcing effect Effects 0.000 description 2
- 230000003252 repetitive effect Effects 0.000 description 2
- 235000009413 Ratibida columnifera Nutrition 0.000 description 1
- 241000510442 Ratibida peduncularis Species 0.000 description 1
- 238000007630 basic procedure Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
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
- G06T5/10—Image enhancement or restoration using non-spatial domain filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- 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]
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention discloses the image enchancing methods under the conditions of a kind of complex illumination, it mainly include initial phase, enhanced fuzzy stage, the image reconstruction stage, downhole video Image Acquisition is completed first, transmission, and wavelet decomposition, low frequency coefficient histogram equalization are carried out to video image, high frequency coefficient carries out coefficient extraction；Then denoising model and new PAL fuzzy enhancement algorithm are designed, realizes the filtering processing to high frequency section and enhanced fuzzy, and Anti-fuzzy processing is carried out to enhanced fuzzy characteristic image；Finally the low frequency coefficient of the high frequency coefficient to Anti-fuzzy processing and histogram equalization carries out wavelet reconstruction, and obtains the effect picture after image enhancement.The present invention can effectively overcome the problems, such as that image characteristic point caused by image shade, light and shade area, half-light, bloom due to caused by underground illumination condition complexity sharply declines, and promote vision facilities further applying in underground coal mine.
Description
Technical field
The invention belongs to field of image recognition, it is related to a kind of image enhancement, Image Denoising Technology, especially by being applied to
The image pre-processing method of underground.
Background technique
The detection for being currently based on the underground coal mine target of image enhancement is of great significance for maintenance mine safety, but works as
Preceding image enhancement technique is in actual complex environment using more difficult.Since underground target stealth capabilities are strong, in underground complexity
It weak output signal that target emanation under environment goes out and is easy to be caused partial target to be mingled in background limit below by background interference
Among clutter.In engineering, the main gain by increasing photodetector, the incident photon-to-electron conversion efficiency for increasing device, increase light
The detectivity that the means such as focal length of lens increase video camera is learned, but the mesh in collected video image under the complex environment of underground
Mark is still very faint, is still difficult to meet underground practical application.
Image enhancement under current complex environment mostly uses single spatial domain enhancing or frequency domain enhancing, and is mainly increasing
Strong picture contrast, prominent image detail, eliminates the research of noise etc..But underground coal mine illumination condition is complicated, exists
The special circumstances such as light is poor, uneven illumination is even, and dust is more, as image shade of the illumination condition caused by complicated, light and shade area, dark
Light, bloom etc., this makes using spatial domain enhancing directly processing grey scale pixel value, and handles frequency domain figure using frequency domain filter
Enhance the requirement of details as the method for realizing enhancing is unable to satisfy downhole video image and should eliminate noise again.Therefore, it studies
A kind of image enchancing method suitable for underground coal mine complex illumination condition is that vision facilities is badly in need of solution in underground coal mine application
Certainly the problem of.
Summary of the invention
Technical problem to be solved by the present invention lies in overcome the deficiencies of the prior art and provide a kind of borehole image enhancing side
Method, solve existing borehole image identification technology be unable to satisfy image shade as underground illumination condition caused by complicated, light and shade area,
Image characteristic point caused by half-light, bloom drastically reduces problem.
The present invention specifically uses following technical scheme to solve above-mentioned technical problem：
1. the image enchancing method under the conditions of a kind of complex illumination, which is characterized in that the image image enchancing method
Including initial phase, enhanced fuzzy stage, image reconstruction stage；Downhole video image can be completed in initial phase to adopt
Collection, transmission, and wavelet decomposition, low frequency coefficient histogram equalization are carried out to video image；In the enhanced fuzzy stage by small wavelength-division
The high frequency coefficient obtained after solution carries out coefficient extraction, fuzzy by the denoising model and new PAL that introduce the fuzzy membership factor
Enhance algorithm, realizes the filtering processing to high frequency section and enhanced fuzzy, make to obtain the fuzzy increasing under different scale, different directions
Strong characteristic image, and Anti-fuzzy processing is carried out to enhanced fuzzy characteristic image；In the height that the image reconstruction stage handles Anti-fuzzy
The low frequency coefficient of frequency coefficient and histogram equalization carries out wavelet reconstruction, and obtains the effect picture after image enhancement.
Further, the initial phase includes the following steps:
(1) visible light camera or infrared camera carry out video image acquisition to underground complex environment；
(2) the collected video image of step 1 is transferred to image procossing by the RJ45 interface or USB interface of video camera
Unit, and carry out video image storage；
(3) video image that step 2 is transferred to image processing unit is subjected to the extraction of video image single frames, by the list of acquisition
Color image frame carries out gray processing processing；
(4) treated the image of gray processing described in step 3 is subjected to multi-scale wavelet decomposition, and obtained under n-th of scale
Low frequency coefficient and 3n direction high frequency coefficient；
(5) low frequency coefficient under obtained n-th of the scale of step 4 is subjected to histogram equalization processing.
Further, the enhanced fuzzy stage includes the following steps:
(1) by the high frequency coefficient in 3n direction obtained after the multi-scale wavelet decomposition, first extraction n=1 when
The high frequency coefficient in 3 directions decomposed, the high frequency coefficient in 3 directions decomposed when then extracting n=2, is then successively extracted
The high frequency coefficient in 3 directions decomposed when n-th of scale；
(2) introducing is successively respectively adopted in the high frequency coefficient in obtain under each scale obtained in step 13 directions
Noise in the denoising model removal image high frequency coefficient of fuzzy membership factor s；
(3) by design fuzzy membership function, the high frequency coefficient after denoising in step 2 is successively become from spatial domain
Change to fuzzy domain set domain；
(4) by design fuzzy enhancement operator, the fuzzy domain set domain that step 3 obtains successively is subjected to the non-linear change of fuzzy set
It changes, i.e., carries out ambiguity function calculating in pixel value of the fuzzy field to high frequency imaging, make to obtain the mould under different scale, different directions
Paste Enhanced feature image；
(5) fuzzy membership function according to designed by step 3, by step 4 treated high frequency coefficient again from fuzzy
Domain set domain transforms to spatial domain, i.e., carries out Anti-fuzzy processing to the enhanced fuzzy characteristic image.
Further, the reconstruction stage includes the following steps:
(1) on n-th of the scale obtained after handling Anti-fuzzy under high frequency coefficient and n-th of scale at histogram equalization
Low frequency coefficient after reason carries out wavelet reconstruction, reconstructs and obscures enhanced image on n-th of scale；
It (2) will be after image pixel matrix after enhanced fuzzy on n-th of scale obtained in step 1 and Anti-fuzzy processing
(n-1)th layer of obtained high frequency coefficient carries out image array construction, and reconstructs the figure after the enhanced fuzzy on (n-1)th scale
Picture；
(3) steps 1 and 2 are repeated, when obtaining with image after the processing of the enhanced fuzzy of the same resolution ratio of original image, are terminated
Repetitive process.
Further, the denoising model, which designs, mainly includes：
It (1), can be according to wavelet decomposition by introducing a kind of fuzzy membership factor s in wavelet threshold denoising model
The noise profile situation in the high frequency coefficient of image adaptively adjusts wavelet threshold, the wavelet threshold in Wavelet Denoising Method model afterwards
Construction of function expression formula is：
In formula, μTFor wavelet threshold function, ωijThe absolute value that (i, j) is put in small echo high frequency coefficient, sgn () are symbol
Number function, s are the fuzzy membership factor, and T is wavelet threshold；
(2) the fixed threshold T in the fuzzy membership factor s substitution wavelet soft-threshold in the model, the factor energy
It is enough adaptively to be adjusted according to the distribution situation of picture noise, the flexibility of model is greatly improved, the calculating of the factor is public
Formula is：
In formula, a is regulatory factor, and a ∈ (0,1], ωijThe absolute value that (i, j) is put in small echo high frequency coefficient, T is small echo
Threshold value；
(3) coefficient of noise information in described image high frequency coefficient differs larger with the coefficient of image information, by right
The high frequency section of multi-scale wavelet decomposition carries out mean square deviation calculating, and obtained mean square deviation will be bigger, then determines some
Coefficient value makes the mean square deviation of all coefficients greater than this numerical value reach minimum, and effect is preferable, then the coefficient value is exactly institute
The threshold value to be chosen, threshold value T calculation formula are：
T=2-n|ωij|σ2/σ'
In formula, n is the highest wavelet decomposition number of plies, and σ is image noise variance, and σ=median (| ωij|)/0.6745, σ '
For image wavelet decomposition coefficient standard deviation, andM is the row maximum value for handling picture, N is
Handle the column maximum value of image.
Further, the PAL fuzzy enhancement algorithm includes：
(1) subordinating degree function μ is designedij：
In formula, ωijThe absolute value that (i, j) is put in small echo high frequency coefficient, ωmaxIt is absolute for greatest coefficient in high frequency section
Value；
ωminFor the minimum absolute coefficient of high frequency section；
(2) fuzzy enhancement operator is designed：
In formula, t is a parameter of control convergence speed, and t is bigger, and convergence rate is faster；Variable element μc∈[0,1]。
Beneficial effects of the present invention are：
Image enchancing method under the conditions of a kind of complex illumination of the present invention, utilizes the multiple dimensioned of wavelet decomposition model
It decomposes, so that the high frequency for realizing processing image respectively is made an uproar by high and low frequency unpack in each resolution ratio after decomposing
Sound and image overall contrast ratio problem；In low frequency part after histogram equalization processing, image grayscale contrast is mentioned
Height, image overall brightness make moderate progress, and gray scale dynamic range expands, and have preferable reinforcing effect.Meanwhile after this algorithm process
Picture low frequency part have no introducing noise signal, algorithm can carry out the profile information of noise signal and image entirety effectively
Separation；By design denoising mathematical model in each resolution ratio of high frequency section, can effectively remove under different scale
High-frequency noise can effectively retain the details of image in enhanced fuzzy image using the nonlinear transformation processing of fuzzy domain set domain
Information and texture feature information, low frequency coefficient retain the general picture property feature of original image.
Image enchancing method under the conditions of a kind of complex illumination of the present invention, the enhanced image of this algorithm with it is original
Image is compared, and the contrast of image significantly improves, and can be good at removing the halation located strongly in comparison of light and shade, image enhancement effect
Fruit naturally, and the Fuzzy Processing function maintain in borehole image the integrity degree of detailed information well and enhance figure
As key feature information, the vision-based detection under underground coal mine complex illumination environment can be applied to.
Detailed description of the invention
Fig. 1 is the initial phase flow chart of the image enhancement under the conditions of a kind of complex illumination of the present invention；
Fig. 2 is the enhanced fuzzy phase flow figure of the image enhancement under the conditions of a kind of complex illumination of the present invention；
Fig. 3 is the reconstruction stage flow chart of the image enhancement under the conditions of a kind of complex illumination of the present invention.
Fig. 4 is the image enhancement processing effect picture under the conditions of a kind of complex illumination of the present invention.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, with reference to the accompanying drawings of the specification to skill of the present invention
Art scheme and specific implementation method are clearly and completely described.
A. a kind of basic procedure of the image enchancing method under the conditions of complex illumination is illustrated referring to Fig.1, it is specific
Steps are as follows:
(1) visible light camera or infrared camera used by Image Acquisition (101) are Explosion-proof Design, the camera shooting
Machine can carry out video image acquisition to the target object in the complex environment of underground；
(2) the collected video image of step 1 is passed through the RJ45 interface or USB interface of video camera by image transmitting (102)
Be transferred to image processing unit, image processing unit include DSP picture processing chip, ISP picture processing chip, CPU or
GPU；
(3) image storage (103) stores the video image that step 2 acquisition equipment is transmitted out, image used
Data storage can be used image pick-up card or directly upload to video image data in Cloud Server；
(4) gray processing (104) mentions the video image progress video image single frames that step 2 is transferred to image processing unit
It takes, and carries out image preservation, gray processing processing is carried out to the single frames color image obtained after preservation；
(5) wavelet decomposition (105) is by treated the corresponding wavelet basis function of image selection of gray processing described in step 4, and
The value of the scale n of wavelet decomposition is designed, n usually takes 1,2,3,4, the ash after wavelet basis function and scale n has been determined
It spends image and carries out wavelet decomposition, the low frequency coefficient under n-th of scale and the high frequency coefficient in 3n direction can be obtained；The small echo
Basic function mainly include Haar wavelet basis Biorthogonal (biorNr.Nd) wavelet systems, db wavelets base, Coiflet
(coifN) wavelet systems, SymletsA (symN) wavelet systems, Molet (morl) small echo, Mexican Hat (mexh) small echo,
Meyer small echo；
(6) low frequency coefficient under obtained n-th of the scale of step 5 is passed through wavelet low frequency by low frequency coefficient extraction (106)
Function is obtained to directly acquire；
(7) low frequency processing (107) carries out histogram equalization processing to the image low frequency coefficient of acquisition, makes low frequency coefficient figure
As pixel is evenly distributed in a certain image, to enhance picture contrast and brightness.
B. referring to Fig. 2 to the enhanced fuzzy stage of the image enchancing method under the conditions of a kind of complex illumination of the present invention
It is illustrated, specific step is as follows:
(1) high frequency coefficient extracts (201) for the high frequency system in 3n direction obtained after the multi-scale wavelet decomposition
The high frequency coefficient in 3 directions decomposed when number, first extraction n=1, the high frequency system in 3 directions decomposed when then extracting n=2
Number, the high frequency coefficient in 3 directions decomposed when then successively extracting n-th of scale；
(2) denoising model designs (202) for the high frequency system in 3 directions obtained under each scale obtained in step 1
Number successively carries out denoising using the Wavelet Denoising Method model for introducing fuzzy membership factor s, small in Wavelet Denoising Method model
Wave threshold function table constructs expression formula：
In formula, μTFor wavelet threshold function, ωijThe absolute value that (i, j) is put in small echo high frequency coefficient, sgn () are symbol
Number function, s are the fuzzy membership factor, and T is wavelet threshold；In fuzzy membership factor s substitution wavelet soft-threshold in model
Fixed threshold T, the factor can adaptively adjust according to the distribution situation of picture noise, and the spirit of model is greatly improved
The calculation formula of activity, the factor is：
In formula, a is regulatory factor, and a ∈ (0,1], ωijThe absolute value that (i, j) is put in small echo high frequency coefficient, T is small echo
Threshold value；The coefficient of noise information in described image high frequency coefficient differs larger with the coefficient of image information, by multiple dimensioned
The high frequency section of wavelet decomposition carries out mean square deviation calculating, and obtained mean square deviation will be bigger, then determines some coefficient value,
The mean square deviation of all coefficients greater than this numerical value is set to reach minimum, and effect is preferable, then the coefficient value is exactly to be chosen
Threshold value, threshold value T calculation formula is：
T=2-n|ωij|σ2/σ'
In formula, n is the highest wavelet decomposition number of plies, and σ is image noise variance, and σ=median (| ωij|)/0.6745, σ '
For
Image wavelet decomposition coefficient standard deviation, expression formula are：
In formula, M is the row maximum value for handling picture, N is the column maximum value for handling image, according to designed Wavelet Denoising Method
Function carries out Wavelet Denoising Method to the high frequency coefficient under each scale；
(3) design subordinating degree function (203) passes through design fuzzy membership function μij, expression formula is：
In formula, ωijThe absolute value that (i, j) is put in small echo high frequency coefficient, ωmaxIt is absolute for greatest coefficient in high frequency section
Value；ωminFor the minimum absolute coefficient of high frequency section；High frequency coefficient after denoising in step 2 is successively subordinate to using fuzzy
Category degree function muij, and then realize that high frequency coefficient transforms from a spatial domain to fuzzy domain set domain；
(4) design fuzzy enhancement operator (204) passes through design fuzzy enhancement operator T (μij), expression formula is：
In formula, t is a parameter of control convergence speed, and t is bigger, and convergence rate is faster；Variable element μc∈[0,1]；
The fuzzy domain set domain that step 3 obtains successively is subjected to fuzzy set nonlinear transformation processing, i.e., in fuzzy field to the picture of high frequency imaging
Element value carries out fuzzy enhancement operator calculating, and then obtains the enhanced fuzzy characteristic image under different scale, different directions；
(5) Anti-fuzzy handles (205) fuzzy membership function according to designed by step 3, by step 4 treated high frequency
Coefficient transforms to spatial domain from fuzzy domain set domain again, i.e., carries out Anti-fuzzy processing to the enhanced fuzzy characteristic image.
C. referring to Fig. 3 to the image reconstruction stage of the image enchancing method under the conditions of a kind of complex illumination of the present invention
It is illustrated, specific implementation step is as follows:
(1) low frequency coefficient and enhanced fuzzy phase process after n-th layer wavelet coefficient construction (301) handles initial phase
N-th layer high frequency coefficient afterwards carries out image array construction；
(2) the image array input wavelet reconstruction function that step 1 constructs is reconstructed in n-th layer wavelet reconstruction (302), and
It reconstructs and obscures enhanced image on n-th of scale, the row and column of the matrix is the image resolution ratio after reconstruct；
(3) m layers of wavelet coefficient construction (303) will obscure enhanced figure on n-th of scale obtained in step 2
As (n-1)th layer of high frequency coefficient progress image array construction after corresponding pixel value and enhanced fuzzy phase process, constructed
The row and column of image array out is 2 times of step 1 image array row and column；
The image array input wavelet reconstruction function that step 3 constructs is reconstructed in (4) m layers of wavelet reconstruction (304), and
It reconstructs and obscures enhanced image on (n-1)th scale, the image resolution ratio after this reconstruct is n-th layer weight structure Larva resolution ratio
2 times；
(5) judge whether m is 0 (305), if m is not 0, returns to (303) and repeat the above steps (3) and step (4), if m
It is 0, then terminates repetitive process, and exports and the enhanced fuzzy of original image equal resolution treated reinforcing effect image.
Claims (6)
1. the image enchancing method under the conditions of a kind of complex illumination, which is characterized in that the image image enchancing method includes
Initial phase, enhanced fuzzy stage, image reconstruction stage；Downhole video Image Acquisition can be completed in initial phase, is passed
It is defeated, and wavelet decomposition, low frequency coefficient histogram equalization are carried out to video image；After the enhanced fuzzy stage is by wavelet decomposition
The high frequency coefficient arrived carries out coefficient extraction, is calculated by the denoising model and new PAL enhanced fuzzy that introduce the fuzzy membership factor
Method realizes the filtering processing to high frequency section and enhanced fuzzy, obtains the enhanced fuzzy characteristic pattern under different scale, different directions
Picture, and Anti-fuzzy processing is carried out to enhanced fuzzy characteristic image；Anti-fuzzy is handled in the image reconstruction stage high frequency coefficient and
The low frequency coefficient of histogram equalization carries out wavelet reconstruction, and obtains the effect picture after image enhancement.
2. the image enchancing method under the conditions of a kind of complex illumination as described in claim 1, which is characterized in that described is initial
The change stage includes the following steps:
1) visible light camera or infrared camera carry out video image acquisition to underground complex environment；
2) the collected video image of step 1 is transferred to image processing unit by the RJ45 interface or USB interface of video camera,
And carry out video image storage；
3) video image that step 2 is transferred to image processing unit is subjected to the extraction of video image single frames, the single frames of acquisition is color
Chromatic graph picture carries out gray processing processing；
4) treated the image of gray processing described in step 3 is subjected to multi-scale wavelet decomposition, and obtains the low frequency under n-th of scale
The high frequency coefficient of coefficient and 3n direction；
5) low frequency coefficient under obtained n-th of the scale of step 4 is subjected to histogram equalization processing.
3. the image enchancing method under the conditions of a kind of complex illumination as described in claim 1, which is characterized in that described is fuzzy
The enhancing stage includes the following steps:
1) it is decomposed when by the high frequency coefficient in 3n direction obtained after the multi-scale wavelet decomposition, first extraction n=1
The high frequency coefficient in 3 directions, the high frequency coefficient in 3 directions decomposed when then extracting n=2, then successively extracts n-th of ruler
The high frequency coefficient in 3 directions decomposed when spending；
2) it is fuzzy that the high frequency coefficient in obtain under each scale obtained in step 13 directions is successively respectively adopted to introducing
Noise in the denoising model removal image high frequency coefficient of degree of membership factor s；
3) by design fuzzy membership function, the high frequency coefficient after denoising in step 2 is successively transformed from a spatial domain to
Fuzzy domain set domain；
4) by design fuzzy enhancement operator, the fuzzy domain set domain that step 3 is obtained successively carries out fuzzy set nonlinear transformation, i.e.,
Ambiguity function calculating is carried out in pixel value of the fuzzy field to high frequency imaging, makes to obtain the fuzzy increasing under different scale, different directions
Strong characteristic image；
5) fuzzy membership function according to designed by step 3 becomes step 4 treated high frequency coefficient from fuzzy domain set domain again
Spatial domain is changed to, i.e., Anti-fuzzy processing is carried out to the enhanced fuzzy characteristic image.
4. the image enchancing method under the conditions of a kind of complex illumination as described in claim 1, it is characterised in that：The image
Reconstruction stage includes the following steps:
1) on n-th of the scale obtained after handling Anti-fuzzy under high frequency coefficient and n-th of scale after histogram equalization processing
Low frequency coefficient carry out wavelet reconstruction, reconstruct and obscure enhanced image on n-th of scale；
2) it will be obtained after image pixel matrix after enhanced fuzzy on n-th of scale obtained in step 1 and Anti-fuzzy processing
(n-1)th layer of high frequency coefficient carry out image array construction, and reconstruct the image after the enhanced fuzzy on (n-1)th scale；
3) steps 1 and 2 are repeated, when obtaining with image after the processing of the enhanced fuzzy of the same resolution ratio of original image, end was repeated
Journey.
5. the image enchancing method under the conditions of a kind of complex illumination as described in claim 1, it is characterised in that：The denoising
Modelling mainly includes：
It 1), can be according to image after wavelet decomposition by introducing a kind of fuzzy membership factor s in wavelet threshold denoising model
High frequency coefficient in noise profile situation adaptively adjust wavelet threshold, the wavelet threshold function structure in Wavelet Denoising Method model
Making expression formula is：
In formula, μTFor wavelet threshold function, ωijThe absolute value that (i, j) is put in small echo high frequency coefficient, sgn () are symbol letter
Number, s are the fuzzy membership factor, and T is wavelet threshold；
2) the fixed threshold T in the fuzzy membership factor s substitution wavelet soft-threshold in the model, and can be made an uproar according to image
The distribution situation of sound adaptively adjusts, and the flexibility of model is greatly improved, and the calculation formula of the factor is：
In formula, a is regulatory factor, and a ∈ (0,1], ωijThe absolute value that (i, j) is put in small echo high frequency coefficient, T is wavelet threshold；
3) coefficient of noise information in described image high frequency coefficient differs larger with the coefficient of image information, by multiple dimensioned
The high frequency section of wavelet decomposition carries out mean square deviation calculating, and obtained mean square deviation will be bigger, then determines some coefficient value,
The mean square deviation of all coefficients greater than this numerical value is set to reach minimum, and effect is preferable, then the coefficient value is exactly to be chosen
Threshold value, threshold value T calculation formula is：
T=2-n|ωij|σ2/σ'
In formula, n is the highest wavelet decomposition number of plies, and σ is image noise variance, and σ=median (| ωij|)/0.6745, σ ' be figure
As coefficient of wavelet decomposition standard deviation, andM is the row maximum value for handling picture, N is processing
The column maximum value of image.
6. the image enchancing method under the conditions of a kind of underground coal mine complex illumination as described in claim 1, it is characterised in that：Institute
The PAL fuzzy enhancement algorithm stated includes：
1) subordinating degree function μ is designedij：
In formula, ωijThe absolute value that (i, j) is put in small echo high frequency coefficient, ωmaxFor greatest coefficient absolute value in high frequency section；
ωminFor the minimum absolute coefficient of high frequency section；
2) fuzzy enhancement operator is designed：
In formula, t is a parameter of control convergence speed, and t is bigger, and convergence rate is faster；Variable element μc∈[0,1]。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810647661.1A CN108876741B (en) | 2018-06-22 | 2018-06-22 | Image enhancement method under complex illumination condition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810647661.1A CN108876741B (en) | 2018-06-22 | 2018-06-22 | Image enhancement method under complex illumination condition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108876741A true CN108876741A (en) | 2018-11-23 |
| CN108876741B CN108876741B (en) | 2021-08-24 |
Family
ID=64340849
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810647661.1A Active CN108876741B (en) | 2018-06-22 | 2018-06-22 | Image enhancement method under complex illumination condition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108876741B (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109658354A (en) * | 2018-12-20 | 2019-04-19 | 上海联影医疗科技有限公司 | A kind of image enchancing method and system |
| CN110187093A (en) * | 2019-06-26 | 2019-08-30 | 上海艾瑞德生物科技有限公司 | A kind of cross-platform detection device detecting fluorescence immunoassay and colloid gold reagent item |
| CN111325680A (en) * | 2020-01-08 | 2020-06-23 | 深圳深知未来智能有限公司 | Image brightening method capable of inhibiting local overexposure |
| CN112669222A (en) * | 2019-10-15 | 2021-04-16 | 国际关系学院 | Digital image defogging method based on wavelet domain histogram equalization |
| CN112862716A (en) * | 2021-02-07 | 2021-05-28 | 江苏科技大学 | Sonar image target enhancement method based on dual-tree complex wavelet and fuzzy theory |
| CN113112435A (en) * | 2020-04-23 | 2021-07-13 | 江苏理工学院 | Variable contrast enhancement method and device for wavelet domain positive and negative image fusion |
| CN113128511A (en) * | 2021-03-31 | 2021-07-16 | 武汉钢铁有限公司 | Coke tissue identification method and device |
| CN113212307A (en) * | 2021-05-28 | 2021-08-06 | 深圳天联星科技有限公司 | Electronic rearview mirror with 360-degree environment monitoring function |
| CN113246859A (en) * | 2021-05-28 | 2021-08-13 | 深圳天联星科技有限公司 | Electronic rearview mirror with driving auxiliary system warning function |
| CN113656462A (en) * | 2021-08-18 | 2021-11-16 | 北京奥康达体育产业股份有限公司 | Wisdom sports park data analysis system based on thing networking |
| US11475569B2 (en) | 2018-05-30 | 2022-10-18 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for image processing |
| CN117422656A (en) * | 2023-12-14 | 2024-01-19 | 西南交通大学 | Low-illumination fuzzy traffic image enhancement method, device, equipment and medium |
| CN118657917A (en) * | 2024-08-19 | 2024-09-17 | 山东济矿鲁能煤电股份有限公司阳城煤矿 | Image enhancement method for underground low-voltage leakage protection of coal mine |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5799112A (en) * | 1996-08-30 | 1998-08-25 | Xerox Corporation | Method and apparatus for wavelet-based universal halftone image unscreening |
| CN101201937A (en) * | 2007-09-18 | 2008-06-18 | 上海医疗器械厂有限公司 | Digital image enhancement method and device based on wavelet restruction and decompose |
| CN101430759A (en) * | 2008-12-04 | 2009-05-13 | 上海大学 | Optimized recognition pretreatment method for human face |
| CN103310423A (en) * | 2013-07-08 | 2013-09-18 | 中国矿业大学(北京) | Mine image intensification method |
| CN103886344A (en) * | 2014-04-14 | 2014-06-25 | 西安科技大学 | Image type fire flame identification method |
| CN103927752A (en) * | 2014-04-21 | 2014-07-16 | 国家电网公司 | Method for monitoring operating state of electrical equipment |
| CN105844601A (en) * | 2016-05-20 | 2016-08-10 | 中国矿业大学(北京) | Mine image enhancement method based on bilateral filtering and multi-scale Retinex algorithm |
-
2018
- 2018-06-22 CN CN201810647661.1A patent/CN108876741B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5799112A (en) * | 1996-08-30 | 1998-08-25 | Xerox Corporation | Method and apparatus for wavelet-based universal halftone image unscreening |
| CN101201937A (en) * | 2007-09-18 | 2008-06-18 | 上海医疗器械厂有限公司 | Digital image enhancement method and device based on wavelet restruction and decompose |
| CN101430759A (en) * | 2008-12-04 | 2009-05-13 | 上海大学 | Optimized recognition pretreatment method for human face |
| CN103310423A (en) * | 2013-07-08 | 2013-09-18 | 中国矿业大学(北京) | Mine image intensification method |
| CN103886344A (en) * | 2014-04-14 | 2014-06-25 | 西安科技大学 | Image type fire flame identification method |
| CN103927752A (en) * | 2014-04-21 | 2014-07-16 | 国家电网公司 | Method for monitoring operating state of electrical equipment |
| CN105844601A (en) * | 2016-05-20 | 2016-08-10 | 中国矿业大学(北京) | Mine image enhancement method based on bilateral filtering and multi-scale Retinex algorithm |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11475569B2 (en) | 2018-05-30 | 2022-10-18 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for image processing |
| US12100157B2 (en) | 2018-05-30 | 2024-09-24 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for image processing |
| CN109658354A (en) * | 2018-12-20 | 2019-04-19 | 上海联影医疗科技有限公司 | A kind of image enchancing method and system |
| CN110187093A (en) * | 2019-06-26 | 2019-08-30 | 上海艾瑞德生物科技有限公司 | A kind of cross-platform detection device detecting fluorescence immunoassay and colloid gold reagent item |
| CN112669222A (en) * | 2019-10-15 | 2021-04-16 | 国际关系学院 | Digital image defogging method based on wavelet domain histogram equalization |
| CN111325680A (en) * | 2020-01-08 | 2020-06-23 | 深圳深知未来智能有限公司 | Image brightening method capable of inhibiting local overexposure |
| CN111325680B (en) * | 2020-01-08 | 2023-04-25 | 深圳深知未来智能有限公司 | Image brightening method capable of inhibiting local overexposure |
| CN113112435A (en) * | 2020-04-23 | 2021-07-13 | 江苏理工学院 | Variable contrast enhancement method and device for wavelet domain positive and negative image fusion |
| CN113112435B (en) * | 2020-04-23 | 2023-09-22 | 江苏理工学院 | Variable contrast enhancement method and device for wavelet domain positive and negative image fusion |
| CN112862716A (en) * | 2021-02-07 | 2021-05-28 | 江苏科技大学 | Sonar image target enhancement method based on dual-tree complex wavelet and fuzzy theory |
| CN113128511B (en) * | 2021-03-31 | 2023-07-25 | 武汉钢铁有限公司 | Coke tissue identification method and device |
| CN113128511A (en) * | 2021-03-31 | 2021-07-16 | 武汉钢铁有限公司 | Coke tissue identification method and device |
| CN113246859A (en) * | 2021-05-28 | 2021-08-13 | 深圳天联星科技有限公司 | Electronic rearview mirror with driving auxiliary system warning function |
| CN113212307A (en) * | 2021-05-28 | 2021-08-06 | 深圳天联星科技有限公司 | Electronic rearview mirror with 360-degree environment monitoring function |
| CN113656462A (en) * | 2021-08-18 | 2021-11-16 | 北京奥康达体育产业股份有限公司 | Wisdom sports park data analysis system based on thing networking |
| CN117422656A (en) * | 2023-12-14 | 2024-01-19 | 西南交通大学 | Low-illumination fuzzy traffic image enhancement method, device, equipment and medium |
| CN117422656B (en) * | 2023-12-14 | 2024-04-12 | 西南交通大学 | Low-illumination fuzzy traffic image enhancement method, device, equipment and medium |
| CN118657917A (en) * | 2024-08-19 | 2024-09-17 | 山东济矿鲁能煤电股份有限公司阳城煤矿 | Image enhancement method for underground low-voltage leakage protection of coal mine |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108876741B (en) | 2021-08-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108876741A (en) | A kind of image enchancing method under the conditions of complex illumination | |
| Yeh et al. | Multi-scale deep residual learning-based single image haze removal via image decomposition | |
| CN103530853B (en) | Infrared light intensity image and infrared polarization figure image intensifying fusion method | |
| CN108875645B (en) | Face recognition method under complex illumination condition of underground coal mine | |
| CN109801250A (en) | Infrared and visible light image fusion method based on ADC-SCM and low-rank matrix expression | |
| CN113421200A (en) | Image fusion method based on multi-scale transformation and pulse coupling neural network | |
| CN109308691A (en) | Infrared and visible light image fusion method based on image enhancement and NSCT | |
| CN110097617B (en) | Image fusion method based on convolutional neural network and significance weight | |
| Luo et al. | Infrared and visible image fusion based on visibility enhancement and hybrid multiscale decomposition | |
| CN108665484A (en) | A kind of dangerous source discrimination and system based on deep learning | |
| CN110751667B (en) | Method for detecting infrared dim and small targets under complex background based on human visual system | |
| Cheng et al. | Adaptive fusion framework of infrared and visual image using saliency detection and improved dual-channel PCNN in the LNSST domain | |
| CN106485684B (en) | A kind of single image based on dual-tree complex wavelet transform goes cloud and mist method | |
| CN1254770C (en) | Image merging method based on maximum expectation value and discrete wavelet frame | |
| Yang et al. | Infrared and visible image fusion based on QNSCT and Guided Filter | |
| Zhang et al. | A novel method of medical image enhancement based on wavelet decomposition | |
| Yang et al. | A novel single image dehazing method | |
| Pei et al. | Application of wavelet-based image fusion in image enhancement | |
| Jia et al. | A wavelet-based approach to improve foggy image clarity | |
| CN109242798A (en) | A kind of Poisson denoising method based on three cross-talk network representations | |
| Kannadhasan et al. | Research issues on digital image processing for various applications in this world | |
| CN107644406A (en) | A kind of image de-noising method based on improved orthogonal matching pursuit | |
| Balamurugan et al. | Classification of Land Cover in Satellite Image using supervised and unsupervised Techniques | |
| Fan | New Denoising Algorithm in Crop Image Processing | |
| Han et al. | Spatial images feature extraction based on bayesian nonlocal means filter and improved contourlet transform |
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 |