CN109658363A - The multilayer sub-block overlapping histogram equalizing method and system that sub-block adaptively merges - Google Patents
The multilayer sub-block overlapping histogram equalizing method and system that sub-block adaptively merges Download PDFInfo
- Publication number
- CN109658363A CN109658363A CN201811228052.9A CN201811228052A CN109658363A CN 109658363 A CN109658363 A CN 109658363A CN 201811228052 A CN201811228052 A CN 201811228052A CN 109658363 A CN109658363 A CN 109658363A
- Authority
- CN
- China
- Prior art keywords
- block
- sub
- algorithm
- ohe
- fast
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 71
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 79
- 230000000670 limiting effect Effects 0.000 claims abstract description 7
- 230000006870 function Effects 0.000 claims description 22
- 230000004927 fusion Effects 0.000 claims description 13
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 3
- 238000012986 modification Methods 0.000 claims description 3
- 230000004048 modification Effects 0.000 claims description 3
- 238000007493 shaping process Methods 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 abstract description 21
- 208000003028 Stuttering Diseases 0.000 abstract description 3
- 230000003321 amplification Effects 0.000 abstract description 3
- 238000003199 nucleic acid amplification method Methods 0.000 abstract description 3
- 230000001186 cumulative effect Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 4
- 238000009826 distribution Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000003384 imaging method Methods 0.000 description 3
- 238000010008 shearing Methods 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 238000005267 amalgamation Methods 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000002156 mixing Methods 0.000 description 2
- 230000002829 reductive effect Effects 0.000 description 2
- 230000003313 weakening effect Effects 0.000 description 2
- 238000003556 assay Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000003014 reinforcing effect Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000005303 weighing 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
- G06T5/40—Image enhancement or restoration using histogram techniques
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Facsimile Image Signal Circuits (AREA)
Abstract
The present invention discloses the multilayer sub-block overlapping histogram equalizing method and system that a seed block adaptively merges, by the movement routine for modifying template in traditional BOHE algorithm, make between current sub-block and previous sub-block that there is only the Non-overlapping Domains that a step-length is mobile, forms fast B OHE algorithm；And it is first defined contrast operation before carrying out fast B OHE algorithm to sub-block, according to the upper limit threshold of number of pixels shared by setting gray level, reduce the slope of pixel accumulated probability density function in fast B OHE algorithm；Then Multilevel method is carried out to original image using the fast B OHE algorithm for limiting contrast, to inhibit the amplification of noise and weaken the stutter bands problem as caused by excessively enhancing, the result of Multilevel method is merged as unit of sub-block finally, enhances the detailed information of image.
Description
Technical field
The present invention relates to technical field of image processing, and in particular to the multilayer sub-block overlapping that a seed block adaptively merges is straight
Square figure equalization methods and system.
Background technique
During imaging, due to natural lighting environment, the influence of the various factors such as imaging device causes imaging to be moved back
Change, causes image local details fuzzy, to directly influence engineering staff to the assay of target signature.Improve image pair
It is a kind of widely used picture quality enhancing technology than degree, this technology mainly passes through the gray scale model of object in extension image
It encloses, realizes the raising of contrast, achieve the purpose that details enhances.
Color histogram equilibrium (HE) is a kind of common method for enhancing picture contrast, standby with simple and quick feature
Favored.Its basic thought is by the way that the gray scale of original image is become full scope from some gray scale interval for comparing concentration
Interior is uniformly distributed, and achievees the purpose that improve contrast.It, can be in the hope of this image according to the mathematical derivation of above-mentioned balancing principle
The transfer function of pixel grey scale is its cumulative histogram function (CDF).
After histogram equalization, brightness of image is improved very much, and tonal range also broadens.Color histogram is equal
Weighing apparatus method makees single mapping to entire image, has expanded the contrast of the close quarters of histogram distribution, has improved figure on the whole
Image contrast, but the contrast of sparse region is had compressed, cause some gray levels to be lost, causes background detail fuzzy.
Histogram equalization is distributed to each regional area of image by sub-block overlapping histogram equalization method (BOHE), according to image
Local characteristics calculate greyscale transformation function, compensate for the problem of gray scale caused by color histogram equilibrium is lost.Pass through a mould
Plate defines the range of image region, using the CDF of template overlay area at this time executes histogram to the center pixel in the region
Figure is balanced, and template moves a pixel every time, aforesaid operations is executed since first pixel of image, to the last a picture
Element completes histogram equalization.
But BOHE the problem of excessively being enhanced with too long execution time and serious contrast, is without by using.Sub-block
On the basis of BOHE, the moving step length by increasing template greatly reduces the histogram equalization method that partly overlaps (POSHE)
Time of algorithm operation.In addition, POSHE carries out histogram equalization to all pixels of template overlay area every time, and will
Gray value after each pixel is balanced every time adds up, then divided by its cumulative frequency is exactly that the POSHE finally obtained enhances
Result.However, the mode of this self-defining moving step length for various sizes of image selection when there is no one both
Fixed standard, step-length is too small, and long operational time, step-length is big, and some images is caused apparent blocking artifact occur.
Based on a kind of innovatory algorithm that the modified local histogram equalization method (ABMHE) of neighboring sub-patch is to POSHE, more
It targetedly selects regional area and carries out equilibrium, the details reinforcing effect of this algorithm is better than POSHE, but needs prior root
According to the active regions and not active region of the ratio identification image of the gradient value numerical values recited of image.Keep the Double-histogram of brightness
Original image histogram is decomposed into two sub- histograms according to the brightness of original image and does histogram equalization respectively by equalization (BBHE)
Processing.When the distribution asymmetry of pixel in image, this method is for keeping the effect of original image image brightness with regard to not ideal enough.It is based on
The Texture similarity equalization (DOTHE) of dominance direction divides the image into texture region and smooth region is individually handled, only to line
Histogram equalization processing is done in reason region, solves the problems, such as that background and object are fuzzy.Similar to BBHE and DOTHE, these are intended to
The contrast enhancement technique for protecting original image image brightness, the contrast effect for enhancing low-light (level) image are not very significant.
Multilayer sub-block is overlapped histogram equalization (MLBOHE), by way of to local histogram is counted in traditional BOHE
It is modified with the mode of template movement, has invented a kind of fast B OHE method, be effectively shortened the runing time of BOHE.Together
When, by studying the relationship of template size histogram equalization (BOHE) result Chong Die with sub-block, it is concluded that in MLBOHE
When the template size used is the 1/2 of original image size, the Luminance Distribution of obtained BOHE enhancing image is better than original image；Make
When with 1/4 that template size is original image size, the details of obtained BOHE enhancing image is more.It the use of template size is original image
As size 1/8 when, obtain BOHE enhancing image contour of object it is more prominent.Based on fast B OHE method, MLBOHE can be with
Be divided into three phases to realize: 1. carry out original image fast B OHE processing three times, the template size used three times point respectively
Not Wei full size 1/2,1/4 and 1/8；2. being used the above-mentioned three-layered node fruit handled three times respectively having a size of 3 × 3,5 × 5 and 7
× 7 median filter is denoised；3. three tomographic images after denoising are fused into piece image.The process of fusion is divided into two again
Secondary, fusion for the first time is that the image after denoising is fused into piece image according to corresponding weight respectively, three layers is melted at this time
Close the ratio that weight ratio is equal to the comentropy difference of three tomographic images and original image；Second of fusion is that above-mentioned three tomographic image is fused
Image is merged with original image.
MLBOHE median filter meeting blurred picture details selected to use when inhibiting noise, and carrying out for the first time
The amalgamation mode selected when fusion is merged based on three layers of global image comentropy with the comentropy difference of original image, this
Mode does not account for the details of topography, therefore is unfavorable for the enhancing of last blending image details.
Summary of the invention
In view of this, the present invention provides the multilayer sub-block overlapping histogram equalizing method and be that a seed block adaptively merges
System, can not only effectively inhibit noise, additionally it is possible to which the phenomenon that weakening the artificial artifact of supersaturation has the image of enhancing certainly
Right visual effect.
The multilayer sub-block that one seed block adaptively merges is overlapped histogram equalizing method, and the sub-block adaptively merges more
Straton block be overlapped histogram equalizing method the following steps are included:
The movement routine of template, makes only to deposit between current sub-block and previous sub-block in the traditional BOHE algorithm of S1, modification
In the mobile Non-overlapping Domain of a step-length, fast B OHE algorithm is formed；
S2, restriction contrast method is introduced in fast B OHE algorithm, according to the upper of number of pixels shared by setting gray level
Threshold value is limited, the slope of pixel accumulated probability density function in fast B OHE algorithm is reduced；
S3, Multilevel method is carried out to original image using the fast B OHE algorithm for limiting contrast, and by the knot of Multilevel method
Fruit is merged as unit of sub-block.
The multilayer sub-block that one seed block adaptively merges is overlapped histogram equalization system, and the sub-block adaptively merges more
It includes following functions module that straton block, which is overlapped histogram equalizing method:
Algorithm correction module, for modifying the movement routine of template in traditional BOHE algorithm, make current sub-block with it is previous
There is only the Non-overlapping Domains that a step-length is mobile between a sub-block, form fast B OHE algorithm；
It limits contrast and introduces module, contrast method is limited for introducing in fast B OHE algorithm, according to setting
The upper limit threshold of number of pixels shared by gray level reduces the slope of pixel accumulated probability density function in fast B OHE algorithm；
Layered shaping Fusion Module is more for using the fast B OHE algorithm for introducing restriction contrast to carry out original image
Layer processing, and the result of Multilevel method is merged as unit of sub-block.
The multilayer sub-block overlapping histogram equalizing method and system that sub-block of the present invention adaptively merges, with MLBOHE
Based on, in MLBOHE algorithm as use median filter denoise caused by blurred image problem, propose use
The method of contrast is limited to inhibit noise.Before carrying out fast B OHE algorithm operating to image, restriction comparison is first done to sub-block
It is close to reduce pixel accumulated probability in fast B OHE algorithm according to the upper limit threshold of number of pixels shared by setting gray level for degree operation
Spend the slope of function.Then Multilevel method is carried out to original image using the fast B OHE algorithm for limiting contrast, to inhibit noise
Amplification and weaken the stutter bands problem as caused by excessively enhancing, finally the result of Multilevel method is carried out as unit of sub-block
Fusion, enhances the detailed information of image.
The multilayer sub-block overlapping histogram equalizing method and system that sub-block of the present invention adaptively merges, algorithm structure
Stablize, it is easily operated, can not only effectively defogging, have a sense of hierarchy the image of defogging more, while reducing to the maximum extent
The contrast of low-light (level) image can be improved under the premise of degree enhancing to protrude details, enable the visual effect of the image of enhancing natural and
Clearly.
Detailed description of the invention
Fig. 1 is the flow diagram for the multilayer sub-block overlapping histogram equalizing method that sub-block of the present invention adaptively merges；
Fig. 2 is the movement routine of template in fast B OHE algorithm compared with the movement routine of template in tradition BOHE algorithm
Schematic diagram；
Fig. 3 is the flow diagram of step S2 in Fig. 1；
Fig. 4 is histogram figure shearing result comparison schematic diagram；
Fig. 5 is the flow diagram of step S3 in Fig. 1；
Fig. 6 is the module frame chart for the multilayer sub-block overlapping histogram equalization system that sub-block of the present invention adaptively merges.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated, it should be understood that and the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
The present invention provides the multilayer sub-block overlapping histogram equalizing method that a seed block adaptively merges, as shown in Figure 1, institute
State the multilayer sub-block overlapping histogram equalizing method that sub-block adaptively merges the following steps are included:
The movement routine of template, makes only to deposit between current sub-block and previous sub-block in the traditional BOHE algorithm of S1, modification
In the mobile Non-overlapping Domain of a step-length, fast B OHE algorithm is formed.
Specifically, the fast B OHE algorithm is using the histogram of template first time overlay area as benchmark histogram, and
Be 1 left right to row from row with step-length by template from the first row, and being expert at right with step-length is 1 to down toward the second row, then again from
It is 1 right left to row from row with step-length, and so on, until having traversed whole image.When template moves every time, due to current son
The histogram of region and previous sub-regions overlapping region had generated in previous sub-block, therefore was creating current sub-district
No longer need to recalculate the histogram of overlapping region when the histogram in domain, it is only necessary to it is non-heavy with previous sub-block to calculate current sub-block
The histogram in folded region.
As shown in Fig. 2, the moving process schematic diagram for the BOHE that Fig. 2 is 3 × 3 using template size, traditional BOHE algorithm
Own in the subregion for needing statistical mask to cover every time during template movement, when calculating the histogram of this sub-regions
The frequency that the gray value of pixel occurs.As shown in Fig. 2 (a), when the use of the size of template being 3 × 3, template covering for the first time
Subregion includes a, b, c, d, e, f, g, h, i, and traditional BOHE algorithm needs to count the pixel value and its frequency of entire subregion
To calculate the histogram of current sub-region；One pixel of template movement, at this time template covering area update be b, e, h, c, f,
I, j, k, l, there is still a need for the pixel values of statistics current sub-region and its frequency for traditional BOHE algorithm to calculate current sub-region
Histogram, but fast B OHE algorithm only needs to count j, and the pixel value and its frequency of k, l, i.e. fast B OHE algorithm only need
Count the histogram of current sub-block and previous sub-block not overlapping region.From traditional BOHE algorithm in the moving process of template
Statistical in it can be found that comprising b, c, e, f, h, the histogram in the region of i previous sub-block and current sub-block all
It is calculated, therefore fast B OHE algorithm no longer counts the histogram of current sub-block and previous sub-block overlapping region again.
Template moves on to the right from the left side of each row in traditional BOHE algorithm, until having traversed whole image, such as Fig. 2
(b) shown in.It is 1 left from row with step-length from the first row shown in mobile route such as Fig. 2 (c) of template in fast B OHE algorithm
It is right to row, and being expert at right with step-length is 1 to down toward the second row, then again from being 1 right left to row from row with step-length, and so on,
Until having traversed whole image.
Therefore, improved BOHE algorithm is only needed when counting local histogram by the straight of template first time overlay area
Side's figure is used as benchmark histogram, only needs to calculate current sub-block and previous sub-block not overlapping region when this rear pattern plate moves every time
Histogram.The calculative pixel quantity of local histogram is generated every time to considerably reduce, and algorithm is complicated
Degree is O (n), and the algorithm complexity of tradition BOHE is
O(MN(m+L))(1)
M × N indicates the size of original image in formula (1), and m × n indicates the size of template, and L indicates the number of greyscale levels of image.
It can thus be seen that improved BOHE algorithm significantly reduces the complexity of algorithm, operational efficiency is improved.
S2, restriction contrast method is introduced in fast B OHE algorithm, according to the upper of number of pixels shared by setting gray level
Threshold value is limited, the slope of pixel accumulated probability density function in fast B OHE algorithm is reduced.
In order to inhibit BOHE algorithm enhance image in noise pollution, before carrying out fast B OHE, to sub-block its
(elder generation) is defined contrast operation, inhibits the amplification of noise by the enhancing of limitation contrast.From histogram equalization principle
The gradation conversion function of known pixels is determined by its CDF, and the formula of the gradation conversion function is as follows:
In formula (2), r indicates former ash angle value, and p indicates that probability density function, L indicate the number of greyscale levels of image, and s indicates to turn
Gray value after changing.From in formula (2) it is found that pixel conversion after gray scale and its accumulated probability density (cumulative histogram) at just
Than, therefore the slope for reducing accumulated probability density function can reduce the ratio of contrast enhancing, to realize the limit of contrast
System.
The contrast that limits is the part sheared in histogram more than the threshold value using the threshold value of a predefined,
Specifically, as shown in figure 3, step S2 include it is following step by step:
S21, setting one gray level shared by number of pixels upper limit threshold β, by former histogram h (n) (n=0,1,2,
... L-1.) in be more than that the number of pixels of threshold value beta is updated to β, and will be more than that the number of pixels of threshold value beta adds up, obtain histogram
It is more than the sum of threshold value in figure, is denoted as Excess；
S22, note m=Excess/L；If h (n) (n=0,1,2 ... L-1.) < β-m, h (n)=h (n)+m is enabled,
Excess=Excess-m；
If S23, β-m < h (n) (n=0,1,2 ... L-1.) < β, enable h (n)=β, Excess=Excess+h (n)-β；
S24, continue to be distributed remaining pixel number, if Excess>0 at this time, h (n) (n=0,1,2 ... L-1.)<β, h
(n)=h (n)+1, Excess=Excess-1, until Excess=0.
As shown in Fig. 4 (a), the threshold value of selection is 3 times of average gray number of pixels, and solid line indicates former histogram, dotted line
Histogram after indicating shearing, by limiting the upper limit of shared pixel number in each gray level, to reduce image cumulative histogram
Slope, such as Fig. 4 (b), solid line indicates former cumulative histogram, and dotted line indicates the cumulative histogram after shearing, to reduce pixel
Tonal range after conversion achievees the purpose that weaken contrast.
S3, Multilevel method is carried out to original image using the fast B OHE algorithm for introducing restriction contrast, and will be at multilayer
The result of reason is merged as unit of sub-block.
According to the relationship for template size and BOHE algorithm the enhancing result studied in MLBOHE algorithm, the present invention uses three layers
Fast B OHE algorithm process mode, it is the 1/2 of original image that wherein first layer, which selects template size, and second layer selection template size is
The 1/4 of original image, it is the 1/8 of original image that third layer, which selects template size,；Picture shared by the gray level that corresponding each layer choosing is selected
Prime number purpose upper limit threshold β is respectively 2 times, 3 times, 4 times of sub-block average gray number of pixels.
It, should be respectively with above-mentioned three layers of BOHE calculation before fusion in order to make fused picture quality be optimal effect
Method enhancing image quality be used as primary concern foundation, i.e. three layers of BOHE arithmetic result merge weight respectively with its image matter
It measures directly proportional.Since the purpose of image enchancing method of the present invention is mainly to disclose the details of image concealing, selection
Image quality criteria is image information entropy, and entropy is bigger, and picture material is abundanter.Then the entropy of enhanced image and original image is poor
Bigger, the details of enhancing is more.
Meanwhile to enhance blending image details, the amalgamation mode for inventing use is merged based on sub-block quality.By three
The enhanced image of layer is respectively classified into the consistent sub-block of population size, is merged as unit of sub-block, sub-image comentropy
It is bigger, indicate its enhancing details it is more, therefore in each layer same position sub-block fusion when weight ratio be equal to they with
The ratio of the image information entropy difference of the sub-block of original image same position, is represented by following formula (3-4):
wi1:wi2:...win=(Ei1-Ei0):(Ei2-Ei0):...(Ein-Ei0) (3)
B={ Bi, i=1 ..., m
In formula (3), WinIndicate the weight of i-th of sub-block of n-th layer image, EinIndicate i-th of son of n-th layer image
The entropy of block, EioIndicate the entropy of i-th of sub-block of original image.B in formula (4)iIndicate fused i-th of sub-block, Wi,jIt indicates
The weight of i-th of sub-block in jth tomographic image, Bi,jIndicate i-th of sub-block in jth tomographic image.
Specifically, as shown in figure 5, step S3 include it is following step by step:
S31, three layers of processing are carried out respectively to original image using the fast B OHE algorithm for introducing restriction contrast；
S32, treated three tomographic image is respectively classified into the consistent sub-block of population size, is carried out as unit of sub-block
Correspond fusion.
The multilayer sub-block that sub-block of the present invention adaptively merges is overlapped histogram equalizing method, using MLBOHE as base
Plinth is limited for, as using blurred image problem caused by median filter denoising, proposing to use in MLBOHE algorithm
The method of contrast inhibits noise.Before carrying out fast B OHE algorithm operating to image, sub-block is first done and limits contrast behaviour
Make, according to the upper limit threshold of number of pixels shared by setting gray level, reduces pixel accumulated probability density letter in fast B OHE algorithm
Several slopes.Then Multilevel method is carried out to original image using the fast B OHE algorithm for limiting contrast, to inhibit putting for noise
Big and weakening stutter bands problem as caused by excessively enhancing, is finally melted the result of Multilevel method as unit of sub-block
It closes, enhances the detailed information of image.
The multilayer sub-block that sub-block of the present invention adaptively merges is overlapped histogram equalizing method, and algorithm structure is stablized,
It is easily operated, can not only effectively defogging, have a sense of hierarchy the image of defogging more, while reducing excessively enhancing to the maximum extent
Under the premise of can improve the contrast of low-light (level) image to protrude details, enable the visual effect of the image of enhancing natural and clear,
Realize low-light (level) image and Misty Image brightness raising and local detail protrusion while effectively avoid it excessively and enhance
The problem of.
It is overlapped histogram equalizing method based on the multilayer sub-block that above-mentioned sub-block adaptively merges, the present invention also provides a seeds
The multilayer sub-block of block adaptive fusion is overlapped histogram equalization system, as shown in fig. 6, multilayer that the sub-block adaptively merges
It includes following functions module that block, which is overlapped histogram equalizing method:
Algorithm correction module 10 makes current sub-block with before for modifying the movement routine of template in traditional BOHE algorithm
There is only the Non-overlapping Domains that a step-length is mobile between one sub-block, form fast B OHE algorithm；
Restriction contrast introduces module 20, limits contrast method for introducing in fast B OHE algorithm, is set with basis
Determine the upper limit threshold of number of pixels shared by gray level, reduces the slope of pixel accumulated probability density function in fast B OHE algorithm；
Layered shaping Fusion Module 30, for using the fast B OHE algorithm for limiting contrast to carry out at multilayer original image
Reason, and the result of Multilevel method is merged as unit of sub-block.
Apparatus above embodiment and embodiment of the method are one-to-one, the simple places of Installation practice, referring to method reality
Apply example.
Each embodiment in this specification is described in a progressive manner, the highlights of each of the examples are with other
The difference of embodiment, the same or similar parts in each embodiment may refer to each other.
Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure
And algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software generally describes each exemplary composition and step according to functionality in the above description.This
A little functions are implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Specially
Industry technical staff can use different methods to achieve the described function each specific application, but this realization is not
It should be more than the scope of the present invention.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor
The combination of capable software module or the two is implemented.Software module can be placed in random access memory, memory, read-only memory,
Institute is public in electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technical field
In the storage medium for any other forms known.
The embodiment of the present invention is described with above attached drawing, but the invention is not limited to above-mentioned specific
Embodiment, the above mentioned embodiment is only schematical, rather than restrictive, those skilled in the art
Under the inspiration of the present invention, without breaking away from the scope protected by the purposes and claims of the present invention, it can also make very much
Form, all of these belong to the protection of the present invention.
Claims (10)
1. the multilayer sub-block that a seed block adaptively merges is overlapped histogram equalizing method, which is characterized in that the sub-block is adaptive
Should merge multilayer sub-block overlapping histogram equalizing method the following steps are included:
The movement routine of template in S1, the traditional BOHE algorithm of modification, makes between current sub-block and previous sub-block that there is only one
The mobile Non-overlapping Domain of a step-length, forms fast B OHE algorithm；
S2, restriction contrast method is introduced in fast B OHE algorithm, according to the upper limit threshold of number of pixels shared by setting gray level
Value reduces the slope of pixel accumulated probability density function in fast B OHE algorithm；
S3, using limit contrast fast B OHE algorithm to original image carry out Multilevel method, and by the result of Multilevel method with
Sub-block is that unit is merged.
2. the multilayer sub-block that sub-block adaptively merges according to claim 1 is overlapped histogram equalizing method, which is characterized in that
The fast B OHE algorithm is moved using the histogram of template first time overlay area as benchmark histogram in this rear pattern plate every time
When dynamic, it is only necessary to calculate the histogram of the Non-overlapping Domain of current sub-block and previous sub-block to create the histogram of current sub-region
Figure.
3. the multilayer sub-block that sub-block adaptively merges according to claim 1 is overlapped histogram equalizing method, which is characterized in that
In the fast B OHE algorithm, it is 1 left right to row from row with step-length, and the right side of being expert at that the movement routine of template, which is from the first row,
It is 1 to down toward the second row with step-length, then again from being 1 right left to row from row with step-length, and so on, until traversing complete figure
Picture.
4. the multilayer sub-block that sub-block adaptively merges according to claim 1 is overlapped histogram equalizing method, which is characterized in that
Calculated and learnt according to the gradation conversion function in fast B OHE algorithm: the gray scale and its accumulated probability density after pixel conversion are at just
Than.
5. the multilayer sub-block that sub-block adaptively merges according to claim 4 is overlapped histogram equalizing method, which is characterized in that
The formula of the gradation conversion function is as follows:
In above formula, after r indicates that former ash angle value, p indicate that probability density function, L indicate that the number of greyscale levels of image, s indicate conversion
Gray value.
6. the multilayer sub-block that sub-block adaptively merges according to claim 1 is overlapped histogram equalizing method, which is characterized in that
Step S2 include it is following step by step:
S21, setting one gray level shared by number of pixels upper limit threshold β, by former histogram h (n) (n=0,1,2 ... L-
1.) number of pixels in more than threshold value beta is updated to β, and the number of pixels more than threshold value beta is added up, and obtains in histogram
More than the sum of threshold value, it is denoted as Excess；
S22, note m=Excess/L；If h (n) (n=0,1,2 ... L-1.) < β-m, enable h (n)=h (n)+m, Excess=
Excess-m；
If S23, β-m < h (n) (n=0,1,2 ... L-1.) < β, enable h (n)=β, Excess=Excess+h (n)-β；
S24, continue to be distributed remaining pixel number, if Excess>0 at this time, h (n) (n=0,1,2 ... L-1.)<β, h (n)=
H (n)+1, Excess=Excess-1, until Excess=0.
7. the multilayer sub-block that sub-block adaptively merges according to claim 1 is overlapped histogram equalizing method, which is characterized in that
Step S3 include it is following step by step:
S31, three layers of processing are carried out respectively to original image using the fast B OHE algorithm for introducing restriction contrast；
S32, treated three tomographic image is respectively classified into the consistent sub-block of population size, is carried out one by one as unit of sub-block
Corresponding fusion.
8. the multilayer sub-block that sub-block adaptively merges according to claim 7 is overlapped histogram equalizing method, which is characterized in that
In three layers of processing, it is the 1/2 of original image that first layer, which selects template size, and it is the 1/ of original image that the second layer, which selects template size,
4, it is the 1/8 of original image that third layer, which selects template size,.
9. the multilayer sub-block that sub-block adaptively merges according to claim 8 is overlapped histogram equalizing method, which is characterized in that
The first layer, the second layer, third layer gray level shared by the upper limit threshold of number of pixels be respectively sub-block average gray pixel
2 times of number, 3 times, 4 times.
10. the multilayer sub-block that a seed block adaptively merges is overlapped histogram equalization system, which is characterized in that the sub-block is adaptive
The multilayer sub-block overlapping histogram equalizing method that should be merged includes following functions module:
Algorithm correction module makes current sub-block and previous height for modifying the movement routine of template in traditional BOHE algorithm
There is only the Non-overlapping Domains that a step-length is mobile between block, form fast B OHE algorithm；
It limits contrast and introduces module, limit contrast method for introducing in fast B OHE algorithm, according to setting gray scale
The upper limit threshold of number of pixels shared by grade reduces the slope of pixel accumulated probability density function in fast B OHE algorithm；
Layered shaping Fusion Module, for using the fast B OHE algorithm for limiting contrast to carry out Multilevel method to original image, and
The result of Multilevel method is merged as unit of sub-block.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811228052.9A CN109658363A (en) | 2018-10-22 | 2018-10-22 | The multilayer sub-block overlapping histogram equalizing method and system that sub-block adaptively merges |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811228052.9A CN109658363A (en) | 2018-10-22 | 2018-10-22 | The multilayer sub-block overlapping histogram equalizing method and system that sub-block adaptively merges |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109658363A true CN109658363A (en) | 2019-04-19 |
Family
ID=66110704
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811228052.9A Pending CN109658363A (en) | 2018-10-22 | 2018-10-22 | The multilayer sub-block overlapping histogram equalizing method and system that sub-block adaptively merges |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109658363A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110276717A (en) * | 2019-06-26 | 2019-09-24 | 纳米视觉(成都)科技有限公司 | A kind of joining method and terminal of image |
| CN111143589A (en) * | 2019-12-06 | 2020-05-12 | Oppo广东移动通信有限公司 | Image processing method and device and storage medium |
| CN111461088A (en) * | 2020-06-17 | 2020-07-28 | 长沙超创电子科技有限公司 | Rail transit obstacle avoidance system based on image processing and target recognition |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6163621A (en) * | 1997-02-27 | 2000-12-19 | Samsung Electronics Co., Ltd | Histogram equalization method and device in contrast enhancement apparatus for image processing system |
| CN101739672A (en) * | 2009-12-02 | 2010-06-16 | 北京中星微电子有限公司 | Method and device for equalizing histogram based on sub-regional interpolation |
-
2018
- 2018-10-22 CN CN201811228052.9A patent/CN109658363A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6163621A (en) * | 1997-02-27 | 2000-12-19 | Samsung Electronics Co., Ltd | Histogram equalization method and device in contrast enhancement apparatus for image processing system |
| CN101739672A (en) * | 2009-12-02 | 2010-06-16 | 北京中星微电子有限公司 | Method and device for equalizing histogram based on sub-regional interpolation |
Non-Patent Citations (3)
| Title |
|---|
| KENTARO KOKUFUTA: "Real-Time Processing of Contrast Limited Adaptive Histogram Equalization on FPGA" * |
| NICHOLAS SIA PIK KONG: "Multiple layers block overlapped histogram equalization for local content emphasis" * |
| 杨光: "限制对比度的多层POSHE自适应图像增强算法" * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110276717A (en) * | 2019-06-26 | 2019-09-24 | 纳米视觉(成都)科技有限公司 | A kind of joining method and terminal of image |
| CN110276717B (en) * | 2019-06-26 | 2023-05-05 | 图码思(成都)科技有限公司 | Image stitching method and terminal |
| CN111143589A (en) * | 2019-12-06 | 2020-05-12 | Oppo广东移动通信有限公司 | Image processing method and device and storage medium |
| CN111461088A (en) * | 2020-06-17 | 2020-07-28 | 长沙超创电子科技有限公司 | Rail transit obstacle avoidance system based on image processing and target recognition |
| CN111461088B (en) * | 2020-06-17 | 2020-09-08 | 长沙超创电子科技有限公司 | Rail transit obstacle avoidance system based on image processing and target recognition |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101505421B (en) | Method of high dynamic range compression with detail preservation and noise constraints | |
| JP5628306B2 (en) | Contrast improvement | |
| CN109658363A (en) | The multilayer sub-block overlapping histogram equalizing method and system that sub-block adaptively merges | |
| CN101783963B (en) | Nighttime image enhancing method with highlight inhibition | |
| JP5031877B2 (en) | Image processing apparatus and image processing method | |
| CN103778599B (en) | A kind of image processing method and system | |
| CN105046658B (en) | A kind of low-light (level) image processing method and device | |
| US9280806B2 (en) | Edge smoothing block filtering and blending | |
| CN107767349B (en) | A kind of method of Image Warping enhancing | |
| CN110766639A (en) | Image enhancement method and device, mobile equipment and computer readable storage medium | |
| JP2010218549A (en) | Method for filtering depth image | |
| WO2019056549A1 (en) | Image enhancement method, and image processing device | |
| CN102231791A (en) | Video image defogging method based on image brightness stratification | |
| JP4490430B2 (en) | A robust recursive envelope operator for fast Retinex processing | |
| CN110400274B (en) | Infrared image enhancement method for vehicle-mounted infrared pedestrian detection | |
| JP6213558B2 (en) | Image processing method and image processing apparatus | |
| CN114092361A (en) | Image local contrast enhancement method for optimizing CLAHE | |
| CN114240767A (en) | Image wide dynamic range processing method and device based on exposure fusion | |
| JP2019121053A (en) | Image processing device and image processing method | |
| JP2002247361A5 (en) | ||
| CN117611501A (en) | Low-illumination image enhancement method, device, equipment and readable storage medium | |
| JPWO2016051716A1 (en) | Image processing method, image processing apparatus, and recording medium for storing image processing program | |
| JP6256704B2 (en) | Image processing method and image processing apparatus | |
| JP2006086788A (en) | Picture correction method, picture correction apparatus, projection type picture display device and brightness unevenness and/or color unevenness correction program | |
| CN115293987A (en) | Improved limited self-adaptive image equalization enhancement algorithm |
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 |