CN104574280B - A kind of two-way histogram equalization processing method of infrared image and system - Google Patents
A kind of two-way histogram equalization processing method of infrared image and system Download PDFInfo
- Publication number
- CN104574280B CN104574280B CN201310480270.2A CN201310480270A CN104574280B CN 104574280 B CN104574280 B CN 104574280B CN 201310480270 A CN201310480270 A CN 201310480270A CN 104574280 B CN104574280 B CN 104574280B
- Authority
- CN
- China
- Prior art keywords
- mrow
- hist
- value
- histogram
- mtd
- 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.)
- Active
Links
Abstract
A kind of two-way histogram equalization processing method of infrared image and system that the present invention is provided, the infrared image formation histogram inputted by handling, and then calculate high, lower boundary, pass through default plateau value and the height feature modeling gradation of image position of centre of gravity, the dark portion and highlights that are divided by the position of centre of gravity are subjected to equilibrium treatment respectively again, so, it is stronger that computational methods are adapted to adaptability, had a clear superiority when handling the image of overall background Small object details, effectively reduce data operation quantity and design cost, simultaneously, the image reform analysis tried to achieve from technical solution of the present invention, infrared image highlights target detail is embodied and be will be apparent from, than through more conforming to eye-observation custom, and relatively existing plateau equalization compares, highlights target detail after technical solution of the present invention processing is unlikely to too to exaggerate, be conducive to the embodiment of infrared image dark portion other details.
Description
Technical field：
The present invention relates to infrared image processing, it particularly relates to infrared image histogram equalization treatment technology.
Background technology：
The infrared image generally existing contrast in the range of Larger Dynamic that thermal infrared imager is obtained is low and details shows not enough
The problem of, common linear stretch or divided linear strength method is made no exception for image background and image detail, is being improved
Background patterns are also improved while picture contrast to occupy display gray scale, and real target details has only occupied few
The display gray scale of amount, causes target detail lack of resolution.Then in thermal infrared imager image procossing, histogram equalization skill
Art particularly Plateau histogram technology is widely applied.The big feature of histogram equalization technology one is the gray scale weight of original image
The heart is necessarily mapped to display gray scale center, and the shortcoming of this histogram equalization when handling overall background image conditions clearly makes
Target detail is obtained to assemble very much, or even not as linear stretch method.Then research and application to plateau equalization is more next
More, the purpose of Plateau histogram seeks to suppress the gray scale proportion that overall background is occupied, and improves the proportion of target detail, but
Plateau histogram very large deviation is possessed, causes the image object detail section after processing excessively to exaggerate, image image reform
Remainder gray level is excessively assembled.Adaptive platform histogram equalization is a kind of method for solving this problem, its embodiment
It is, for different images input gray level, not use unified plateau value, but the universal general character being distributed according to image histogram,
It is many to improve the plateau value for being distributed in two ends using close to " V " type plateau value, reduce the plateau value being centrally located.
But, although adaptive platform histogram equalization technology transfer weighting function substantially substantially increases operand, but
Also this situation without still image gray scale benchmark in the range of Larger Dynamic of unsuitable thermal infrared imager, for thermal infrared imager
The different scenes of acquisition, weighting function will constantly do flexible processing, and algorithm is complex.
The content of the invention：
An object of the present invention be to provide infrared image processing effect preferably, data operation quantity is greatly reduced, saves
The technical scheme of design cost.
An object of the present invention is to solve plateau equalization gradation of image center hardly possible when handling infrared image
To determine, gradation of image centre-of gravity shift is caused to cause image vision photosensitive uncoordinated and the unconspicuous problem of target detail.
Therefore, the present invention provides a kind of two-way histogram equalization processing method of infrared image, including：Inputted by processing
Infrared image formation histogram, corresponding pixel grayscale value interval note is obtained according to the infrared image number of bits b
For n and with it is described it is interval in the corresponding each statistics with histogram value of each pixel grayscale value be designated as HIST_STAT（n）；By the picture
The ascending corresponding each statistics with histogram value that adds up of plain gray-scale value is until when being more than threshold value A, current histogram is counted
The corresponding pixel grayscale value of value is used as lower boundary；By the descending each histogram system for adding up corresponding of the grey scale pixel value
Evaluation is until when being more than the threshold value A, regard pixel grayscale value corresponding to current histogram statistical value as high border；To described
Pixel grayscale between high and low border is worth corresponding statistics with histogram value and carries out the accumulation calculating of platform first, to obtain platform
First accumulation calculating result is designated as HIST_T, and the accumulation calculating of platform first is the Nogata that will be greater than default plateau value T
Figure statistical value replaces with T and added up；Each pixel grayscale value correspondence statistics with histogram value between the height border is made
The accumulation calculating of platform first, until when being more than the 1/2 of the first accumulation calculating result HIST_T, with current histogram statistical value
Corresponding pixel grayscale value is that image reform is designated as HIST_M；Respectively to each pixel grey scale less than the picture center of gravity HIST_M
Level value, each pixel grayscale value more than described image center of gravity HIST_M make equilibrium treatment.
It is described respectively to each pixel grayscale value less than the picture center of gravity HIST_M, more than described image center of gravity HIST_M
Each pixel grayscale value include as equilibrium treatment：
（a）The accumulation calculating of platform second is made to each pixel grayscale value less than image reform HIST_M,
The second accumulation calculating of platform formula is：
The accumulation result of platform second is obtained, HIST_ADD (m) is designated as, m ∈ [0, HIST_M), and obtain and be less than image weight
The platform statistics summation HIST_SUM_LOW=HIST_ADD in heart HIST_M regions（0）；
（b）Make the accumulation calculating of platform the 3rd, the platform to each pixel grayscale value more than image reform HIST_M
Three accumulation calculating formula are：
The accumulation result of platform the 3rd is obtained, HIST_ADD (m), m ∈ (HIST_M, 2 is designated asb- 1], and obtain be more than figure
As each platform statistics summation HIST_SUM_HIGH=HIST_ADD (2 of center of gravityb-1)；
（c）According to formula：
VOUT=128-128*HIST_ADD(m)/HIST_SUM_LOW,m∈[0,HIST_M)
Result of calculation, to less than image reform HIST_M each pixel grayscale value corresponding pixel points quantity make equilibrium at
Reason；
（d）According to formula：
VOUT=128+128*HIST_ADD(m)/HIST_SUM_HIGH,m∈(HIST_M,2b-1]
Result of calculation, to more than image reform HIST_M each pixel grayscale value corresponding pixel points quantity make equilibrium at
Reason.
The step（c）With（d）It is preceding also to include image reform HIST_M being mapped to the step of histogram is shown.
The statistics with histogram value is the pixel quantity of each pixel grayscale value.
Further, the threshold value A is the 1/4 of the pixel sum of the infrared image that is inputted.
The present invention also provides a kind of infrared image two-way histogram equalization processing system, including：Input statistics with histogram list
Member, for by handling inputted infrared image formation histogram, correspondence to be obtained according to the infrared image number of bits b
Pixel grayscale value interval be designated as n and with it is described it is interval in the corresponding each statistics with histogram value of each pixel grayscale value be designated as
HIST_STAT（n）；Height border acquiring unit, for being worth the ascending corresponding each institute that adds up by the pixel grayscale
Statistics with histogram value is stated when more than threshold value A, using pixel grayscale value corresponding to current histogram statistical value as lower boundary,
And for by the descending corresponding each statistics with histogram value that adds up of the grey scale pixel value until more than the threshold
During value A, pixel grayscale value corresponding to current histogram statistical value is regard as high border；Image reform acquiring unit, for pair
The corresponding each statistics with histogram value of each pixel grayscale value between the height border carries out platform and added up, total to obtain
Platform counts accumulated value HIST_T, wherein, the cumulative statistics with histogram for referring to will be greater than default plateau value T of the platform
Value replaces with T and added up, and makees described flat to each pixel grayscale value correspondence statistics with histogram value between the height border
Platform adds up, until when being more than the 1/2 of the first accumulation calculating result HIST_T, with the corresponding pixel of current histogram statistical value
Gray-scale value is image reform HIST_M；Bidirectional equalization processing unit, for respectively to each less than the picture center of gravity HIST_M
Pixel grayscale value, each pixel grayscale value more than described image center of gravity HIST_M make equilibrium treatment.
The bidirectional equalization processing unit includes two-way statistics summing elements, bidirectional equalization output unit, described two-way equal
The processing unit that weighs is used for described respectively to each pixel grayscale value less than the picture center of gravity HIST_M, more than described image center of gravity
HIST_M each pixel grayscale value is referred to as equilibrium treatment：
The two-way statistics summing elements, are used for：
Make the accumulation calculating of platform second to each pixel grayscale value less than image reform HIST_M, the platform second tires out
Plus calculation formula is：
The accumulation result of platform second is obtained, HIST_ADD (m) is designated as, m ∈ [0, HIST_M), and obtain and be less than image weight
The platform statistics summation HIST_SUM_LOW=HIST_ADD in heart HIST_M regions（0）；
Make the accumulation calculating of platform the 3rd to each pixel grayscale value more than image reform HIST_M, the platform the 3rd tires out
Plus calculation formula is：
The accumulation result of platform the 3rd is obtained, HIST_ADD (m), m ∈ (HIST_M, 2 is designated asb- 1], and obtain be more than figure
As each platform statistics summation HIST_SUM_HIGH=HIST_ADD (2 of center of gravityb-1)；
The bidirectional equalization output unit, is used for
According to formula：
VOUT=128-128*HIST_ADD(m)/HIST_SUM_LOW,m∈[0,HIST_M)
Result of calculation, to less than image reform HIST_M each pixel grayscale value corresponding pixel points quantity make equilibrium at
Reason；
According to formula：
VOUT=128+128*HIST_ADD(m)/HIST_SUM_HIGH,m∈(HIST_M,2b-1]
Result of calculation, to more than image reform HIST_M each pixel grayscale value corresponding pixel points quantity make equilibrium at
Reason.
The bidirectional equalization processing unit, which is additionally operable to image reform HIST_M being mapped to the histogram, to be shown.
The statistics with histogram value is the pixel quantity of each pixel grayscale value.
The threshold value A is the 1/4 of the pixel sum of the infrared image that is inputted.
A kind of two-way histogram equalization processing method of infrared image and system that the present invention is provided, are inputted by processing
Infrared image formation histogram, and then high-low threshold value is calculated, gradation of image is calculated by default plateau value and the high-low threshold value
Position of centre of gravity, then the dark portion and highlights that are divided by the position of centre of gravity are subjected to equilibrium treatment respectively, in this way, computational methods are adapted to
Adaptability is stronger, is had a clear superiority when handling the image of overall background Small object details, effectively reduces data operation quantity and sets
Count cost.
It is most of for dark background with image, in case of target detail is in highlights, tried to achieve from technical solution of the present invention
Image reform analysis, compared with classical histogram equalization, to infrared image highlights target detail embody will be apparent from, compared with
Classics histogram equalizing method the inventive method more conform to eye-observation custom, and relatively existing plateau equalization
Compare, technical solution of the present invention processing after highlights target detail be unlikely to too to exaggerate, be conducive to infrared image dark portion its
The embodiment of his details.
Meanwhile, technical scheme is from image reform position, to the picture of the dark portion less than the position of centre of gravity
Plain gray-scale value, the both direction more than the pixel grayscale value of the highlights of the position of centre of gravity use different balanced proportions
Do image equalization processing and be conducive to the control to image sequence position of centre of gravity, image reform is more stablized, anti-noise jamming ability
It is stronger, show as before and after image sequence that light and shade is basically identical between frame data, be conducive to avoid image sequence light and shade from flashing to show
As.
Brief description of the drawings：
Fig. 1 is a kind of step flow chart of the two-way histogram equalization processing method of infrared image of the present invention.
Fig. 2 is a kind of structure chart of the two-way histogram equalization processing system of infrared image of the present invention.
Fig. 3 a to 3c are a kind of image reform of the two-way histogram equalization processing method of infrared image and system of the invention
Principle analysis figure.
Embodiment：
As shown in figure 1, the present invention provides kind of the two-way histogram equalization processing method of infrared image:Including：
S1：The infrared image formation histogram inputted by handling, is obtained according to the infrared image number of bits b
Corresponding pixel grayscale value interval is designated as n and each statistics with histogram value corresponding with each pixel grayscale value in the interval
It is designated as HIST_STAT（n）.In the present embodiment, the histogram is using gray level as transverse axis, using pixel count value as the longitudinal axis, when
So in other embodiments can also pixel probability of occurrence be longitudinal axis etc., the simply difference of calculation, this is not done to be existing
Repeat, the number of bits b units are bit, for example, can be 2bit, 4bit, 8bit, 16bit etc., n is pixel grey scale area
Between, by taking the infrared input pictures of 14bit as an example, n ∈ [0,214-1]。
S2：It is worth the ascending corresponding each statistics with histogram value that adds up until more than threshold value A by the pixel grayscale
When, it regard pixel grayscale value corresponding to current histogram statistical value as lower boundary；
S3：By the descending corresponding each statistics with histogram value that adds up of the grey scale pixel value until more than the threshold value
During A, pixel grayscale value corresponding to current histogram statistical value is regard as high border；
S4：It is worth corresponding statistics with histogram value progress platform first to the pixel grayscale between the high and low border to tire out
Plus calculate, HIST_T is designated as to obtain platform the first accumulation calculating result, the accumulation calculating of platform first is that will be greater than presetting
The plateau value T statistics with histogram value replaces with T and added up；
S5：The cumulative meter of platform first is made to each pixel grayscale value correspondence statistics with histogram value between the height border
Calculate, until when being more than the 1/2 of the first accumulation calculating result HIST_T, with the corresponding pixel grey scale of current histogram statistical value
Level value is that image reform is designated as HIST_M；
S6：Respectively to each pixel grayscale value less than the picture center of gravity HIST_M, more than described image center of gravity HIST_M
Each pixel grayscale value make equilibrium treatment.
In the present embodiment, the step S6：Respectively to each pixel grayscale value less than the picture center of gravity HIST_M, big
Include in described image center of gravity HIST_M each pixel grayscale value as equilibrium treatment：
（a）Make the accumulation calculating of platform second, the platform to each pixel grayscale value less than image reform HIST_M
Two accumulation calculating formula are：
The accumulation result of platform second is obtained, HIST_ADD (m) is designated as, m ∈ [0, HIST_M), and obtain and be less than image weight
The platform statistics summation HIST_SUM_LOW=HIST_ADD in heart HIST_M regions（0）；
（b）Make the accumulation calculating of platform the 3rd, the platform to each pixel grayscale value more than image reform HIST_M
Three accumulation calculating formula are：
The accumulation result of platform the 3rd is obtained, HIST_ADD (m), m ∈ (HIST_M, 2 is designated asb- 1], and obtain be more than figure
As each platform statistics summation HIST_SUM_HIGH=HIST_ADD (2 of center of gravityb-1)；
（c）According to formula：
VOUT=128-128*HIST_ADD(m)/HIST_SUM_LOW,m∈[0,HIST_M)
Result of calculation, to less than image reform HIST_M each pixel grayscale value corresponding pixel points quantity make equilibrium at
Reason；
（d）According to formula：
VOUT=128+128*HIST_ADD(m)/HIST_SUM_HIGH,m∈(HIST_M,2b-1]
Result of calculation, to more than image reform HIST_M each pixel grayscale value corresponding pixel points quantity make equilibrium at
Reason.
In the present embodiment, the step (c),（d）It is preceding also include step image reform HIST_M is mapped to it is described
The step of histogram is shown, this is exported for the ease of image equalization processing, certainly not necessarily.
It is another it should be noted that, the step（a）、（b）It is respectively to more than image reform and less than the pixel of image reform
Gray-scale value region is carried out respectively, can in no particular order, and step（c）、（d）Be also it is similar do equilibrium treatment respectively,
Equally can in no particular order, also, in other embodiments, can also be for example in no particular order respectively to more than image reform picture
Plain gray-scale value region carries out cumulative and then equilibrium treatment, to carrying out adding up then less than image reform pixel grayscale value region
In equilibrium treatment, the present embodiment due to be will the step (c),（d）It is preceding to have added the aobvious processing of center of gravity figure therefore so arrangement,
But it is not to be limited with the present embodiment.
In the present embodiment, the statistics with histogram value is the pixel quantity of each pixel grayscale value, and through reality
Test, the threshold value A for the pixel sum of the infrared image that is inputted 1/4 when, it is relatively more reasonable.
As shown in Fig. 2 the present invention also provides a kind of infrared image two-way histogram equalization processing system 1, including：
Histogram statistical unit 11 is inputted, for by handling inputted infrared image formation histogram, according to described
Infrared image number of bits b come obtain corresponding pixel grayscale value interval be designated as n and with it is described it is interval in each pixel grayscale
It is worth corresponding each statistics with histogram value and is designated as HIST_STAT（n）.
Height border acquiring unit 12 is each described straight corresponding to ascending add up for being worth by the pixel grayscale
Square figure statistical value is when more than threshold value A, using pixel grayscale value corresponding to current histogram statistical value as lower boundary, and
For by the descending corresponding each statistics with histogram value that adds up of the grey scale pixel value until more than the threshold value A
When, it regard pixel grayscale value corresponding to current histogram statistical value as high border.
Image reform acquiring unit 13, for corresponding each described to each pixel grayscale value between the height border
Statistics with histogram value carries out platform and added up, to obtain total platform statistics accumulated value HIST_T, wherein, the platform is cumulative to be referred to
The statistics with histogram value that will be greater than default plateau value T replaces with T and added up, to each pixel between the height border
Gray-scale value correspondence statistics with histogram value is made the platform and added up, until more than the 1/ of the first accumulation calculating result HIST_T
When 2, using the corresponding pixel grayscale value of current histogram statistical value as image reform HIST_M.
Bidirectional equalization processing unit 14, for respectively to each pixel grayscale value less than the picture center of gravity HIST_M, big
Make equilibrium treatment in described image center of gravity HIST_M each pixel grayscale value.
The bidirectional equalization processing unit 14 includes two-way statistics summing elements 141, bidirectional equalization output unit 142, comes
Complete respectively to each pixel grayscale value less than the picture center of gravity HIST_M, each pixel more than described image center of gravity HIST_M
Gray-scale value makees equilibrium treatment.
The two-way statistics summing elements 141, are used for：
Make the accumulation calculating of platform second to each pixel grayscale value less than image reform HIST_M, the platform second tires out
Plus calculation formula is：
The accumulation result of platform second is obtained, HIST_ADD (m) is designated as, m ∈ [0, HIST_M), and obtain and be less than image weight
The platform statistics summation HIST_SUM_LOW=HIST_ADD in heart HIST_M regions（0）；
Make the accumulation calculating of platform the 3rd to each pixel grayscale value more than image reform HIST_M, the platform the 3rd tires out
Plus calculation formula is：
The accumulation result of platform the 3rd is obtained, HIST_ADD (m), m ∈ (HIST_M, 2 is designated asb- 1], and obtain be more than figure
As each platform statistics summation HIST_SUM_HIGH=HIST_ADD (2 of center of gravityb-1)。
Hold, the bidirectional equalization output unit 142 is used for：
According to formula：
VOUT=128-128*HIST_ADD(m)/HIST_SUM_LOW,m∈[0,HIST_M)
Result of calculation, to less than image reform HIST_M each pixel grayscale value corresponding pixel points quantity make equilibrium at
Reason；
According to formula：
VOUT=128+128*HIST_ADD(m)/HIST_SUM_HIGH,m∈(HIST_M,2b-1]
Result of calculation, to more than image reform HIST_M each pixel grayscale value corresponding pixel points quantity make equilibrium at
Reason.
In the present embodiment, the bidirectional equalization processing unit 14 is additionally operable to image reform HIST_M being mapped to described straight
Square figure is shown.
In the present embodiment, the statistics with histogram value is the pixel quantity of each pixel grayscale value, the threshold
Value A is the 1/4 of the pixel sum of the infrared image that is inputted.
The two-way histogram equalization processing system 1 of a kind of infrared image provided by the present invention, its principle and the infrared figure
As two-way histogram equalization processing method is essentially identical, therefore do not repeat separately.
As shown in Fig. 3 a to 3c, show respectively classical histogram equalization image procossing, plateau equalization processing and
The histogram of equilibrium treatment of the present invention is shown, and display image center of gravity.
The bi-directional platform that the present invention of Fig. 3 a histogram equalization, Fig. 3 b plateau equalization and Fig. 3 c is designed
Histogram equalizing method, is all the center that resulting image reform position is mapped to display gray scale, institute of the present invention
The suitable adaptability of the gradation of image center of gravity calculation method of design is stronger, has when handling the image of overall background Small object details bright
Aobvious advantage.
It is most of for dark background with image, in case of target detail is in highlights, tried to achieve from technical solution of the present invention
Image reform analysis, compared with the classical histogram equalization shown in Fig. 3 a, to infrared image highlights target detail embody will
It is more obvious, more conform to eye-observation custom than classical histogram equalizing method the inventive method；With putting down shown in Fig. 3 b
Platform histogram equalization compares, and highlights target detail is unlikely to too to exaggerate, and is conducive to the body of infrared image dark portion other details
It is existing.
In addition, balancing procedure of the present invention is from image reform position, toward the dark portion less than the position and more than the position
Highlights both direction do image equalization processing using different balanced proportions, be conducive to the control to image sequence position of centre of gravity
System, image reform is more stablized, and anti-noise jamming ability is stronger, shows as before and after image sequence light and shade basic one between frame data
Cause, be conducive to the phenomenon for avoiding image sequence light and shade from flashing.
It should be appreciated that one of ordinary skill in the art just can be according to the design of present patent application without creative work
Make many modifications and variations.Therefore, all those skilled in the art according to the design of this patent on the basis of prior art
It is upper to pass through the available technical scheme of logical analysis, reasoning, or a limited experiment, it should all protected determined by this patent
In the range of shield.
Claims (8)
1. a kind of two-way histogram equalization processing method of infrared image, it is characterised in that including：
The infrared image formation histogram inputted by handling, it is corresponding to obtain according to the infrared image number of bits b
Pixel grayscale value interval is designated as n and each statistics with histogram value corresponding with each pixel grayscale value in the interval is designated as
HIST_STAT(n)；
It is worth ascending cumulative corresponding each statistics with histogram value when more than threshold value A by the pixel grayscale, ought
Pixel grayscale value corresponding to preceding statistics with histogram value is used as lower boundary；
It is worth descending cumulative corresponding each statistics with histogram value when more than the threshold value A by the pixel grayscale,
It regard pixel grayscale value corresponding to current histogram statistical value as high border；
Corresponding statistics with histogram value is worth to the pixel grayscale between the height border and carries out the accumulation calculating of platform first, with
Obtain platform the first accumulation calculating result and be designated as HIST_T, the accumulation calculating of platform first is that will be greater than default plateau value T
The statistics with histogram value replaces with T and added up；
The accumulation calculating of platform first is made to each pixel grayscale value correspondence statistics with histogram value between the height border, until
More than the first accumulation calculating result HIST_T 1/2 when, using the corresponding pixel grayscale value of current histogram statistical value as
Image reform is designated as HIST_M；
Respectively to each pixel grayscale value less than described image center of gravity HIST_M, each picture more than described image center of gravity HIST_M
Plain gray-scale value makees equilibrium treatment：
(a) accumulation calculating of platform second is made to each pixel grayscale value less than image reform HIST_M, the platform second tires out
Plus calculation formula is：
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>A</mi>
<mi>D</mi>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>A</mi>
<mi>D</mi>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>S</mi>
<mi>T</mi>
<mi>A</mi>
<mi>T</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>S</mi>
<mi>T</mi>
<mi>A</mi>
<mi>T</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo><</mo>
<mi>T</mi>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>A</mi>
<mi>D</mi>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>T</mi>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>o</mi>
<mi>t</mi>
<mi>h</mi>
<mi>e</mi>
<mi>r</mi>
<mi>w</mi>
<mi>i</mi>
<mi>s</mi>
<mi>e</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>,</mo>
</mrow>
The accumulation result of platform second is obtained, HIST_ADD (m) is designated as, m ∈ [0, HIST_M), and obtain and be less than image reform area
The platform statistics total value HIST_SUM_LOW=HIST_ADD (0) in domain；
(b) accumulation calculating of platform the 3rd is made to each pixel grayscale value more than image reform HIST_M, the platform the 3rd tires out
Plus calculation formula is：
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>A</mi>
<mi>D</mi>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>A</mi>
<mi>D</mi>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>S</mi>
<mi>T</mi>
<mi>A</mi>
<mi>T</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>S</mi>
<mi>T</mi>
<mi>A</mi>
<mi>T</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo><</mo>
<mi>T</mi>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>A</mi>
<mi>D</mi>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>T</mi>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>o</mi>
<mi>t</mi>
<mi>h</mi>
<mi>e</mi>
<mi>r</mi>
<mi>w</mi>
<mi>i</mi>
<mi>s</mi>
<mi>e</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>,</mo>
</mrow>
The accumulation result of platform the 3rd is obtained, HIST_ADD (m), m ∈ (HIST_M, 2 is designated asb- 1], and obtain be more than image reform
The platform statistics total value HIST_SUM_HIGH=HIST_ADD (2 in regionb-1)；
(c) according to formula：
VOUT=128-128*HIST_ADD (m)/HIST_SUM_LOW, m ∈ [0, HIST_M)
Result of calculation, equilibrium treatment is made to each pixel grayscale value corresponding pixel points quantity less than image reform HIST_M；
(d) according to formula：
VOUT=128+128*HIST_ADD (m)/HIST_SUM_HIGH, m ∈ (HIST_M, 2b-1]
Result of calculation, equilibrium treatment is made to each pixel grayscale value corresponding pixel points quantity more than image reform HIST_M.
2. the two-way histogram equalization processing method of infrared image as claimed in claim 1, it is characterised in that in the step
(c) and before (d) also include image reform HIST_M being mapped to the step of histogram is shown.
3. the two-way histogram equalization processing method of infrared image as described in any one in claim 1 to 2, its feature exists
In the statistics with histogram value is the pixel quantity of each pixel grayscale value.
4. the two-way histogram equalization processing method of infrared image as claimed in claim 3, it is characterised in that the threshold value A is
The 1/4 of the pixel sum of the infrared image inputted.
5. a kind of two-way histogram equalization processing system of infrared image, it is characterised in that including：
Histogram statistical unit is inputted, for by handling inputted infrared image formation histogram, according to the infrared figure
Corresponding pixel grayscale value interval is obtained as number of bits b and is designated as n and corresponding with each pixel grayscale value in the interval
Each statistics with histogram value be designated as HIST_STAT (n)；
Height border acquiring unit, unites for being worth ascending cumulative corresponding each histogram by the pixel grayscale
Evaluation until more than threshold value A when, using pixel grayscale value corresponding to current histogram statistical value as lower boundary, and for by
The grey scale pixel value is descending to add up corresponding each statistics with histogram value when more than the threshold value A, ought
Pixel grayscale value corresponding to preceding statistics with histogram value is used as high border；
Image reform acquiring unit, for being worth corresponding each histogram to each pixel grayscale between the height border
Statistical value progress platform adds up, to obtain total platform statistics accumulated value, wherein, the platform is cumulative to be referred to will be greater than default put down
The platform value T statistics with histogram value replaces with T and added up, to each pixel grayscale value correspondence between the height border
Statistics with histogram value is made the platform and added up, until when being more than the 1/2 of the first accumulation calculating result HIST_T, with current histogram
The corresponding pixel grayscale value of statistical value is image reform HIST_M；
Bidirectional equalization processing unit, for respectively to each pixel grayscale value less than the picture center of gravity HIST_M, more than described
Image reform HIST_M each pixel grayscale value makees equilibrium treatment；
Wherein, the bidirectional equalization processing unit includes two-way statistics summing elements, bidirectional equalization output unit, described two-way equal
The processing unit that weighs is used for described respectively to each pixel grayscale value less than the picture center of gravity HIST_M, more than described image center of gravity
HIST_M each pixel grayscale value is referred to as equilibrium treatment：The two-way statistics summing elements, are used for：
Make the accumulation calculating of platform second, the cumulative meter of the platform second to each pixel grayscale value less than image reform HIST_M
Calculating formula is：
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>A</mi>
<mi>D</mi>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>A</mi>
<mi>D</mi>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>S</mi>
<mi>T</mi>
<mi>A</mi>
<mi>T</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>S</mi>
<mi>T</mi>
<mi>A</mi>
<mi>T</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo><</mo>
<mi>T</mi>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>A</mi>
<mi>D</mi>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>T</mi>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>o</mi>
<mi>t</mi>
<mi>h</mi>
<mi>e</mi>
<mi>r</mi>
<mi>w</mi>
<mi>i</mi>
<mi>s</mi>
<mi>e</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>,</mo>
</mrow>
The accumulation result of platform second is obtained, HIST_ADD (m) is designated as, m ∈ [0, HIST_M), and obtain and be less than image reform
The platform statistics summation HIST_SUM_LOW=HIST_ADD (0) in HIST_M regions；
Make the accumulation calculating of platform the 3rd, the cumulative meter of the platform the 3rd to each pixel grayscale value more than image reform HIST_M
Calculating formula is：
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>A</mi>
<mi>D</mi>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>A</mi>
<mi>D</mi>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>S</mi>
<mi>T</mi>
<mi>A</mi>
<mi>T</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>S</mi>
<mi>T</mi>
<mi>A</mi>
<mi>T</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo><</mo>
<mi>T</mi>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>S</mi>
<mi>T</mi>
<mo>_</mo>
<mi>A</mi>
<mi>D</mi>
<mi>D</mi>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>T</mi>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>o</mi>
<mi>t</mi>
<mi>h</mi>
<mi>e</mi>
<mi>r</mi>
<mi>w</mi>
<mi>i</mi>
<mi>s</mi>
<mi>e</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>,</mo>
</mrow>
The accumulation result of platform the 3rd is obtained, HIST_ADD (m), m ∈ (HIST_M, 2 is designated asb- 1], and obtain be more than image reform
Each platform statistics summation HIST_SUM_HIGH=HIST_ADD (2b-1)；
The bidirectional equalization output unit, is used for
According to formula：
VOUT=128-128*HIST_ADD (m)/HIST_SUM_LOW, m ∈ [0, HIST_M)
Result of calculation, equilibrium treatment is made to each pixel grayscale value corresponding pixel points quantity less than image reform HIST_M；
According to formula：
VOUT=128+128*HIST_ADD (m)/HIST_SUM_HIGH, m ∈ (HIST_M, 2b-1]
Result of calculation, equilibrium treatment is made to each pixel grayscale value corresponding pixel points quantity more than image reform HIST_M.
6. the two-way histogram equalization processing system of infrared image as claimed in claim 5, it is characterised in that the bidirectional equalization
Processing unit, which is additionally operable to image reform HIST_M being mapped to the histogram, to be shown.
7. the two-way histogram equalization processing system of infrared image as described in any one in claim 5 to 6, its feature exists
In the statistics with histogram value is the pixel quantity of each pixel grayscale value.
8. the two-way histogram equalization processing system of infrared image as claimed in claim 7, it is characterised in that the threshold value A is
The 1/4 of the pixel sum of the infrared image inputted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310480270.2A CN104574280B (en) | 2013-10-15 | 2013-10-15 | A kind of two-way histogram equalization processing method of infrared image and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310480270.2A CN104574280B (en) | 2013-10-15 | 2013-10-15 | A kind of two-way histogram equalization processing method of infrared image and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104574280A CN104574280A (en) | 2015-04-29 |
| CN104574280B true CN104574280B (en) | 2017-08-25 |
Family
ID=53090264
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310480270.2A Active CN104574280B (en) | 2013-10-15 | 2013-10-15 | A kind of two-way histogram equalization processing method of infrared image and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104574280B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110223248B (en) * | 2019-05-22 | 2021-03-02 | 浙江黑卡电气有限公司 | Dynamic range compression method for infrared image detail enhancement |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101340511A (en) * | 2008-08-07 | 2009-01-07 | 中兴通讯股份有限公司 | Adaptive video image enhancing method based on lightness detection |
| CN101719989A (en) * | 2009-11-30 | 2010-06-02 | 北京中星微电子有限公司 | Method and system for backlight compensation |
| CN102693531A (en) * | 2012-01-11 | 2012-09-26 | 河南科技大学 | Adaptive double-platform based infrared image enhancement method |
| US8311327B2 (en) * | 2009-08-24 | 2012-11-13 | Kabushiki Kaisha Toshiba | System and method for backlit image detection |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8218868B2 (en) * | 2007-06-13 | 2012-07-10 | Sensors Unlimited, Inc. | Method and apparatus for enhancing images |
| US20110026818A1 (en) * | 2009-07-30 | 2011-02-03 | Jonathan Yen | System and method for correction of backlit face images |
-
2013
- 2013-10-15 CN CN201310480270.2A patent/CN104574280B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101340511A (en) * | 2008-08-07 | 2009-01-07 | 中兴通讯股份有限公司 | Adaptive video image enhancing method based on lightness detection |
| US8311327B2 (en) * | 2009-08-24 | 2012-11-13 | Kabushiki Kaisha Toshiba | System and method for backlit image detection |
| CN101719989A (en) * | 2009-11-30 | 2010-06-02 | 北京中星微电子有限公司 | Method and system for backlight compensation |
| CN102693531A (en) * | 2012-01-11 | 2012-09-26 | 河南科技大学 | Adaptive double-platform based infrared image enhancement method |
Non-Patent Citations (4)
| Title |
|---|
| Contrast Enhancement Using Brightness Preserving Bi-Histogram Equalization;YEONG-TAEKGI M 等;《IEEE transactions on Consumer Electronics》;19971231;第43卷(第1期);第3节 * |
| 基于双平台直方图的红外图像增强算法;宋岩峰 等;《红外与激光工程》;20080430;第37卷(第2期);308-311 * |
| 基于平台直方图的红外图像自适应增强算法;王炳健等;《光子学报》;20050228;第24卷(第2期);第1.1节 * |
| 红外图像直方图双向均衡技术研究;陈钱 等;《红外与毫米波学报》;20031231;第22卷(第6期);328-430 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104574280A (en) | 2015-04-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103854259B (en) | Image processing equipment and the method for handling image | |
| CN104134191B (en) | Image de-noising method and its device | |
| CN104636118B (en) | QR Quick Response Code self-adaption binaryzation treating method and apparatus based on photo-equilibrium | |
| CN105184754B (en) | Method for enhancing picture contrast | |
| CN105046658B (en) | A kind of low-light (level) image processing method and device | |
| CN110555808B (en) | Image processing method, device, equipment and machine-readable storage medium | |
| CN107948514A (en) | Image virtualization processing method, device and mobile equipment | |
| CN102930517A (en) | Image enhancement method implemented through histogram equalization | |
| CN110428389B (en) | Low-light-level image enhancement method based on MSR theory and exposure fusion | |
| CN102096909A (en) | Improved unsharp masking image reinforcing method based on logarithm image processing model | |
| CN107563984A (en) | A kind of image enchancing method and computer-readable recording medium | |
| CN105027161B (en) | Image processing method and image processing equipment | |
| CN101527038A (en) | Improved method for enhancing picture contrast based on histogram | |
| CN107203980B (en) | Underwater target detection image enhancement method of self-adaptive multi-scale dark channel prior | |
| CN102208101A (en) | Self-adaptive linearity transformation enhancing method of infrared image | |
| CN113962859B (en) | Panorama generation method, device, equipment and medium | |
| EP4007988A1 (en) | Systems and methods for noise reduction | |
| CN104574280B (en) | A kind of two-way histogram equalization processing method of infrared image and system | |
| CN105513015B (en) | A kind of clearness processing method of Misty Image | |
| CN110400274A (en) | A kind of vehicle mounted infrared pedestrian detection infrared image enhancing method | |
| CN104751441A (en) | Method and device for estimating image noise | |
| CN105608685B (en) | A kind of secondary histogram equalization image enchancing method and system of histogram modification | |
| CN107533540A (en) | The average computation device of nonlinear data | |
| CN110009574A (en) | A kind of method that brightness, color adaptively inversely generate high dynamic range images with details low dynamic range echograms abundant | |
| CN104299195B (en) | A kind of image defogging method based on auto-adaptive doublethreshold and dark channel prior |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address |
Address after: 214135 -6, Linghu Avenue, Wuxi Taihu international science and Technology Park, Wuxi, Jiangsu, China, 180 Patentee after: China Resources micro integrated circuit (Wuxi) Co., Ltd Address before: No.180-22, Linghu Avenue, Taihu International Science and Technology Park, Wuxi, Jiangsu, 214135 Patentee before: WUXI CHINA RESOURCES SEMICO Co.,Ltd. |
|
| CP03 | Change of name, title or address |