CN104537615B - A kind of local Retinex Enhancement Methods based on HSV color spaces - Google Patents

A kind of local Retinex Enhancement Methods based on HSV color spaces Download PDF

Info

Publication number
CN104537615B
CN104537615B CN201410736010.1A CN201410736010A CN104537615B CN 104537615 B CN104537615 B CN 104537615B CN 201410736010 A CN201410736010 A CN 201410736010A CN 104537615 B CN104537615 B CN 104537615B
Authority
CN
China
Prior art keywords
mrow
image
subgraph
pixel
mfrac
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
Application number
CN201410736010.1A
Other languages
Chinese (zh)
Other versions
CN104537615A (en
Inventor
王洪玉
曹静
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Dalian University of Technology
Original Assignee
Dalian University of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Dalian University of Technology filed Critical Dalian University of Technology
Priority to CN201410736010.1A priority Critical patent/CN104537615B/en
Publication of CN104537615A publication Critical patent/CN104537615A/en
Application granted granted Critical
Publication of CN104537615B publication Critical patent/CN104537615B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Abstract

The invention discloses a kind of local Retinex defogging methods based on HSV color spaces, the image pockety that misted to greasy weather situation carries out image enhaucament.The method of the present invention comprises the following steps：A, multi-level Misty Image is handled in different color spaces, realizes picture superposition and definition lifting；B, image size is obtained, the size of adaptive judgement sub-image, piecemeal is carried out to the luminance picture of Misty Image；Enhancing processing is carried out to each subgraph of luminance component, improves the definition of image；C, the area ratio factor of each pixel in each overlapping subgraph is calculated, summation is weighted to each overlapping subgraph and realizes image co-registration；D, the saturation degree component of image is handled, reduces image color；The present invention clearly can rapidly reduce the scenery details of different levels, improve the visual effect of image, suitable for the system of taking photo by plane of different zones.

Description

A kind of local Retinex Enhancement Methods based on HSV color spaces
Technical field
The invention belongs to the technical field of Image Information Processing, the present invention is a kind of part based on HSV color spaces Retinex defogging methods, suitable for the atomization image shot under enhancing haze weather.
Background technology
Numerous outdoor monitoring systems of computer application require the feature that can extract image exactly, and in days such as hazes Under vaporous condition, often due in air fine particle scattering process influence, cause obtain image seriously degrade, not only influence The visual effect of image, and disturb the extraction of characteristics of image, make outdoor monitoring system can not normal operation, so as to bringing Serious potential safety hazard.Therefore, make atomization image sharpening using effective and quick defogging method, there is important reality Meaning.
Existing image defogging method is broadly divided into two major classes：Processing method based on image restoration and based on image enhaucament Processing method.
Based on the algorithm of image restoration using atmospherical scattering model as theoretical foundation, atmospherical scattering model describes Misty Image and moved back Change reason；Image Restoration Algorithm passes through inverting image degradation process, restoration scenario albedo, to obtain the fog free images of optimization. For such method from the essence of image degradation, obtained result is typically more natural.But atmospherical scattering model is a morbid state Equation, can not direct solution equation draw the albedo of scene, it is necessary to using constraintss such as prior informations, to atmospheric scattering mould Parameter in type is estimated, and then reduces the albedo of scene.
Algorithm based on image enhaucament does not consider image quality decrease reason, by carrying out overall or local drawing to image Stretch, brightness of image, contrast and prominent details are improved, so as to improve visual effect.At present, the main stream approach based on image enhaucament It is Retinex algorithm, based on the color constancy of people's vision, the luminance component of image is estimated by filtering, obtains reflecting component Realize image defogging.Local detail can be lost but this method malleable color ratio causes color distortion, while stretching, and it is difficult To choose the parameter value of suitable wave filter.
Paper name：A Multiscale Retinex for Bridging the Gap Between Color Images And the Human Observation of Scenes, periodical：IEEE Transaction on image Processing, time：1997, the 7th phase, page 965 to page 976, Daniel J.Jobson et al. proposed band color recovery Multi-Scale Retinex Algorithm, according to color constancy principle, image illumination component, reduction are estimated using Gaussian filter function Image reflecting component, linear stretch mode improve brightness and contrast.But filter parameter chooses difficulty, and can not reduce Multi-level Misty Image.
Proprietary term：Smart Image Enhancement Process, periodical：Patent Application Publication, time：In October, 2010, numbering：US 2010/0266214Al, Daniel J.Jobson et al. is right first The contrast of input picture and brightness carry out judging whether image has mist, then to Misty Image carry out histogram equalization and The enhancing algorithm that MSRCR algorithms are combined, the enhancing that finally iterates is until definition and brightness reach requirement or iteration time Number reaches the upper limit.
Paper name：Nonlinear Transfer Function-Based Local Approach for Color Image Enhancement, periodical：IEEE Transactions on Consumer Electronics, time：2011, 16th phase, page 858 to page 865, it is proposed that a kind of image enchancing method based on HSV color spaces, be first transformed into image HSV color spaces, contrast enhancing individually then is carried out using non-linear transfer function to V component (i.e. luminance component), without Change the S components and H components of image, to retain color, can effectively strengthen the even image of uneven illumination, but be not suitable for Misty Image Processing.
The content of the invention
The present invention is in order to overcome the shortcomings of conventional images defogging technology, there is provided a kind of part based on HSV color spaces Retinex defogging methods, the definition of atomization image can either be strengthened, improve the visual effect of image, there is faster fortune again Scanning frequency degree, real-time is good, wide adaptation range.
A kind of local Retinex defogging methods based on HSV color spaces, this method comprise the following steps：
A, multi-level Misty Image is handled in different color spaces, realizes that picture superposition and definition carry Rise；Rgb color space processing image easily causes color distortion；HSV color spaces are suitable to human visual perception, describe image Saturation degree, tone, the feature of brightness；Saturation degree describes influence of the mist to the Scene colors depth；Brightness describes atmosphere light to image The influence of definition；Tone describes the type of Scene colors；Strengthen image in HSV color spaces, do not influence the color components of RGB tri- Ratio, so as to not interfere with the skew of color.
B, image size is obtained, the size of adaptive judgement sub-image, piecemeal is carried out to the luminance picture of Misty Image； Enhancing processing is carried out to each subgraph of luminance component, improves the definition of image；
There are four kinds of fogless, mist, dense fog, thick fog situations in Misty Image, single Gaussian convolution yardstick can not adapt to a variety of Situation；In order to obtain the details in more thick fog regions, corresponding convolution yardstick is chosen, estimates luminance component using Gaussian convolution, also Former clear reflecting component.
According to the size of source images, determine piecemeal number and piecemeal after subgraph size；According to partly overlapping side The luminance picture (V component) of image is divided into several subgraphs by formula；The Gauss of each subgraph is calculated around convolution function Scale size, MSR enhancing processing is carried out to each subgraph.Comprise the following steps that：
B1：After obtaining the luminance picture of source images, calculate the size of image, and after piecemeal subgraph size；By image It is divided into the indefinite subgraph of number according to the image size of setting, partly overlapping mode；Picture size is big, then the piecemeal chosen Number is more；
B2：Gauss is in convolution function, the effect of image after scale size decision processing：Large scale (such as σ=200) place Color fidelity is good after reason；Dynamic range compression is good after small yardstick (such as σ=15) processing；MSR algorithms are by three yardstick weighted sums Realize color fidelity and dynamic range compression simultaneously；MSR enhancing processing is carried out to subgraph：
Wherein, (i, j) is the card coordinate of pixel, and n is yardstick label, span：[1, N], N are that yardstick is always individual Number, V (i, j) be input luminance picture, Vo(i, j) be output luminance picture, WnFor weights, value is { 1/3,1/3,1/ 3 }, Fn(i, j) is that Gauss is around convolution function, expression formula：
Fn(i, j)=Kexp [- (i2+j2)/σ2] (2)
Wherein, (i, j) is the card coordinate of pixel, and σ is yardstick, and K is constant.
C, the area ratio factor of each pixel in each overlapping subgraph is calculated, each overlapping subgraph is weighted and asked With realize image co-registration；
Luminance picture after fusion treatment is strengthened is carried out to each subgraph；With current pixel point in each subgraph most Neighboring pixel areas is scale factor in proximal border, using the area and the ratio of total overlapping area as weights, is merged using alpha Mode is merged, and eliminates blocking effect；Comprise the following steps that：
C1, multi-Scale Retinex Algorithm determine the value of pixel according to the relativeness of surrounding pixel；One direction is (horizontal Or vertical) choose surrounding neighbors pixel, it is impossible to influence of the reflection border surrounding pixel to current pixel point completely；But by picture The relativeness of all pixels (both horizontally and vertically) together decides on around vegetarian refreshments and border；Therefore, choose horizontal and vertical Direction pixel, the scale factor using area as current sub-image pixels point；
C2, calculate pixel (i, j) neighborhood territory pixel area in each subgraph：With pixel (i, j) for starting point, to nearest Border (both horizontally and vertically) distance for skirt into square region area ωn；Calculate the weight containing current pixel The gross area ω in folded region；With the neighborhood territory pixel area ω of each subgraphnRatio with overlapping region gross area ω is weights：
Wherein, n be subgraph label, span：[1, N], N are the number of overlapping subgraph.
C3, the overlapping subgraph of fusion, obtain clearly reflecting component；Merged using Alpha amalgamation modes, area ratio Value is used as weights：
Wherein, (i, j) be pixel card coordinate, Rv(i, j) be fusion after brightness value, Rn(i, j) is that (i, j) exists End value in each subgraph, n be subgraph label, span：[1, N], N are the number of overlapping subgraph.
D, the saturation degree component of image is handled, reduces image color；
In Misty Image, the presence of mist causes the color saturation of target scene to reduce；S components in HSV color spaces Reflect the saturation degree of each scene point in image；There is corresponding relation in the saturation degree of HSV color spaces and brightness, according to saturation degree The change of pixel saturation degree is judged with the corresponding relation that brightness before and after enhancing changes；Step is as follows：
D1, the conversion formula according to HSV color spaces and rgb color space, saturation degree component (S components) and luminance component Meet following relation between (V component)：
V=max (rgb)
Wherein, S is saturation degree component, and V is luminance component, and rgb is three color components of rgb color space.
D2, image is assumed before and after enhancing, the minimum value of RGB triple channels will not change；Obtain brightness before and after strengthening The changing value of component, calculate the S components after change：
Wherein, S is untreated saturation degree component, and V is non-reinforced luminance component, and S ' is the saturation degree component after processing, V ' is luminance component after enhancing, and K represents the minimum value of RGB triple channels.
D3, in practice, strengthening the minimum value of front and rear RGB triple channels can change；Compensated using Gamma conversion The dynamic change of minimum value.
So=(S ')γ (7)
Wherein, SoFor the saturation degree component of output, S ' is the saturation degree component after handling for the first time, and γ is Gamma coefficients.
The beneficial effects of the invention are as follows：With reference to the advantage of Local treatment image, using the definition of MSR enhancing Misty Images And color rendition, improve visual detail.Blocking effect phenomenon be present for Local treatment image, propose area Alpha fusion methods, Using weight of the ratio of area as subgraph, to remove blocking effect.Using the characteristic of HSV color spaces, Gauss volume is reduced Long-pending execution number, to improve execution efficiency.Using the mutual conversion regime between HSV color spaces and rgb color space, obtain The conversion of saturation degree component and the relation of luminance component are taken, to realize color rendition, reduces image fault.
Brief description of the drawings
Fig. 1 is the particular flow sheet of the present invention.
The V component that Fig. 2 is the present invention strengthens algorithm flow chart.
The S components that Fig. 3 is the present invention strengthen algorithm flow chart.
Fig. 4 is the instance processes result figure of the present invention.Wherein, (a) is original atomization image；(b) it is (a) initial V points Spirogram picture；(c) it is the initial S component images of (a)；(d) it is V component image after the enhancing of (a)；(e) it is the RGB triple channels of (a) Minimum value；(f) it is S component images after the first time enhancing of (a)；(g) it is defogging result after the first time enhancing of (a)；(h) it is (a) S component images after second of enhancing；(i) the defogging result for the present invention to (a)；(j) defogging knots of the MSRCR to (a) Fruit.
Embodiment
The present invention is elaborated with reference to specific embodiments and the drawings.
A, the HSV triple channel components of atomization image (see accompanying drawing 4 (a) Suo Shi) are obtained.V component (see accompanying drawing 4 (b) Suo Shi) table The brightness of diagram picture, scope are [0-255], and 0 represents black, and 255 represent white；S components (see accompanying drawing 4 (c) Suo Shi) represent figure The saturation degree component of picture, scope are display image between [0-1], and each pixel is multiplied by 255.
B, image is strengthened using adaptive local Retienx enhancing algorithms, shown in the schematic flow sheet as accompanying drawing 2 of method：
B1, the number for choosing image block first, small image selection piecemeal number is small, and big image block number is big；The figure As size is：2627*1926mm, it is 10*7 to choose piecemeal number, and subgraph size is 262*275mm；
B2, according to the subgraph size chosen, to the V component (accompanying drawing 4 (b)) of input picture using the piecemeal that partly overlaps Method carry out piecemeal；According to formula (2), Gaussian convolution filtering is carried out to the V component subgraph after piecemeal and obtains luminance component, Luminance component acquisition reflecting component, which is removed, according to formula (1) realizes that brightness strengthens；
C, subgraph is merged using the Alpha fusion methods based on area factor to obtain enhanced V component：
C1, strengthen to obtain each subgraph according to step B2, the influence area of overlapping region pixel is calculated, with the pixel To border (both horizontally and vertically) length and width of nearest distance as area, area is asked for；It is overlapping with four subgraphs Area factor α is asked for according to formula (3) in region as the gross arean.In specific implementation, overlapping region is four points of subgraph One of；
C2, area factor α obtained according to C1nAfterwards, summation is weighted to the pixel of overlapping region according to formula (4), Image after being merged.Thus, obtain shown in enhanced V component such as accompanying drawing 4 (d).
D, strengthen S components, recover the color of image.Wherein, the schematic flow sheet of algorithm is as shown in Figure 3：
D1, according to formula (5) (6), it is (such as attached to obtain the changing value of V component and the minimum value figure of input picture before and after enhancing Shown in Fig. 4 (e))；Then according to the relation between V component before and after S components and RGB minimum values and enhancing, enhanced S points are obtained Amount is as shown in accompanying drawing 4 (f).
After D2, the S components (accompanying drawing 4 (f)) obtained according to D1 strengthen, there is skew (such as accompanying drawing 4 (g) institute in the color of image Show)；Because counter the changing of RGB minimum values that enhancing is front and rear；Therefore, Gamma compensation is carried out according to formula (7).Specific In implementation, Gamma penalty coefficient γ values are 0.8.Obtain shown in enhanced S component maps such as accompanying drawing 4 (h).
By above-mentioned steps, Fig. 4 (i) is defogging result of the present invention to atomization image 4 (a).Fig. 4 (j) is MSRCR methods To Fig. 4 (a) defogging result.

Claims (3)

  1. A kind of 1. local Retinex defogging methods based on HSV color spaces, it is characterised in that following steps：
    A, multi-level Misty Image is transformed into HSV color spaces to be handled, realizes picture superposition and definition Lifting；
    B, image size, the size of adaptively selected sub-image, according to partly overlapping mode to Misty Image are obtained Luminance picture carry out piecemeal；MSR enhancing processing is carried out to each subgraph of luminance picture, improves the definition of image；Specifically Step is as follows：
    B1：After obtaining the luminance picture of source images, the size of image and the size of subgraph after piecemeal are calculated；By image according to setting Fixed image size is divided into the indefinite subgraph of number in a manner of partly overlapping；Picture size is big, then the piecemeal number chosen It is more；
    B2：Gauss is in convolution function, the effect of image after scale size decision processing：Color fidelity after large scale processing It is good；Dynamic range compression is good after small scale processing；Three yardstick weighted sums are realized color fidelity and dynamic model by MSR algorithms simultaneously Confined pressure contracts；MSR enhancing processing is carried out to subgraph：
    <mrow> <msub> <mi>V</mi> <mi>o</mi> </msub> <mrow> <mo>(</mo> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>)</mo> </mrow> <mo>=</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>W</mi> <mi>n</mi> </msub> <mo>{</mo> <mi>l</mi> <mi>o</mi> <mi>g</mi> <mo>&amp;lsqb;</mo> <mi>V</mi> <mrow> <mo>(</mo> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>)</mo> </mrow> <mo>&amp;rsqb;</mo> <mo>-</mo> <mi>l</mi> <mi>o</mi> <mi>g</mi> <mo>&amp;lsqb;</mo> <mi>V</mi> <mrow> <mo>(</mo> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>)</mo> </mrow> <mo>*</mo> <msub> <mi>F</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>)</mo> </mrow> <mo>&amp;rsqb;</mo> <mo>}</mo> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>;</mo> </mrow>
    Wherein, (i, j) is the card coordinate of pixel, and n is yardstick label, span：[1, N], N are yardstick total number, V (i, j) be input luminance picture, Vo(i, j) be output luminance picture, WnFor weights, value is { 1/3,1/3,1/3 }, Fn (i, j) is that Gauss is around convolution function, expression formula：
    Fn(i, j)=Kexp [- (i2+j2)/σ2] (2)
    Wherein, (i, j) is the card coordinate of pixel, and σ is yardstick, and K is constant；
    C, according to the alpha amalgamation modes based on area, weights of each pixel in each overlapping subgraph are calculated, to each Result of the pixel in each overlapping subgraph is weighted summation and realizes image co-registration；
    D, according to the relation between saturation degree component and luminance component, the saturation degree component of image is handled, and uses Gamma is compensated, and reduces image color.
  2. 2. local Retinex defogging methods according to claim 1, it is characterised in that in step C, enter to each subgraph Luminance picture after row fusion treatment is strengthened；With current pixel point neighboring pixel areas in the nearest border in each subgraph For scale factor, using the area and the ratio of total overlapping area as weights, merged using alpha amalgamation modes, eliminate block Effect；Comprise the following steps that：
    C1, multi-Scale Retinex Algorithm determine the value of pixel according to the relativeness of surrounding pixel；Side horizontally or vertically To selected surrounding neighbors pixel, it is impossible to influence of the reflection border surrounding pixel to current pixel point completely；But by pixel The relativeness of all pixels together decides on around point and border both horizontally and vertically；Therefore, horizontal and vertical side is chosen To pixel, the scale factor using area as current sub-image pixels point；
    C2, calculate current pixel point neighborhood territory pixel area in each subgraph：Using current pixel point as starting point, to nearest level With the distance on vertical direction border for skirt into square region area ωn；Calculate the overlapping region containing current pixel Gross area ω；With the neighborhood territory pixel area ω of each subgraphnRatio with overlapping region gross area ω is weights：
    <mrow> <msub> <mi>&amp;alpha;</mi> <mi>n</mi> </msub> <mo>=</mo> <mfrac> <msub> <mi>&amp;omega;</mi> <mi>n</mi> </msub> <mi>&amp;omega;</mi> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>3</mn> <mo>)</mo> </mrow> </mrow>
    Wherein, n be subgraph label, span：[1, N], N are the number of overlapping subgraph；
    C3, the overlapping subgraph of fusion, obtain clearly reflecting component；Merged using Alpha amalgamation modes, area ratio is made For weights：
    <mrow> <msub> <mi>R</mi> <mi>v</mi> </msub> <mrow> <mo>(</mo> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>)</mo> </mrow> <mo>=</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>&amp;alpha;</mi> <mi>n</mi> </msub> <msub> <mi>R</mi> <mi>n</mi> </msub> <mrow> <mo>(</mo> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>4</mn> <mo>)</mo> </mrow> </mrow>
    Wherein, (i, j) be pixel card coordinate, Rv(i, j) be fusion after brightness value, Rn(i, j) is (i, j) in each son End value in image, n be subgraph label, span：[1, N], N are the number of overlapping subgraph.
  3. 3. local Retinex defogging methods according to claim 1 or 2, it is characterised in that in step D, Misty Image In, the presence of mist causes the color saturation of target scene to reduce；S components in HSV color spaces reflect each field in image The saturation degree at sight spot；Be present corresponding relation in the saturation degree of HSV color spaces and brightness, become according to brightness before and after saturation degree and enhancing The corresponding relation of change judges the change of pixel saturation degree；Step is as follows：
    D1, the conversion formula according to HSV color spaces and rgb color space, meet such as between saturation degree component and luminance component Lower relation：
    <mrow> <mtable> <mtr> <mtd> <mrow> <mi>V</mi> <mo>=</mo> <mi>max</mi> <mrow> <mo>(</mo> <mi>r</mi> <mi>g</mi> <mi>b</mi> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mi>S</mi> <mo>=</mo> <mfrac> <mrow> <mi>max</mi> <mrow> <mo>(</mo> <mi>r</mi> <mi>g</mi> <mi>b</mi> <mo>)</mo> </mrow> <mo>-</mo> <mi>m</mi> <mi>i</mi> <mi>n</mi> <mrow> <mo>(</mo> <mi>r</mi> <mi>g</mi> <mi>b</mi> <mo>)</mo> </mrow> </mrow> <mrow> <mi>max</mi> <mrow> <mo>(</mo> <mi>r</mi> <mi>g</mi> <mi>b</mi> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>=</mo> <mn>1</mn> <mo>-</mo> <mfrac> <mrow> <mi>min</mi> <mrow> <mo>(</mo> <mi>r</mi> <mi>g</mi> <mi>b</mi> <mo>)</mo> </mrow> </mrow> <mrow> <mi>max</mi> <mrow> <mo>(</mo> <mi>r</mi> <mi>g</mi> <mi>b</mi> <mo>)</mo> </mrow> </mrow> </mfrac> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>&amp;DoubleRightArrow;</mo> <mi>S</mi> <mo>=</mo> <mn>1</mn> <mo>-</mo> <mfrac> <mrow> <mi>min</mi> <mrow> <mo>(</mo> <mi>r</mi> <mi>g</mi> <mi>b</mi> <mo>)</mo> </mrow> </mrow> <mi>V</mi> </mfrac> </mrow> </mtd> </mtr> </mtable> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>5</mn> <mo>)</mo> </mrow> </mrow>
    Wherein, S is saturation degree component, and V is luminance component, and rgb is three color components of rgb color space；
    D2, image is assumed before and after enhancing, the minimum value of RGB triple channels will not change；Obtain luminance component before and after strengthening Changing value, calculate change after S components：
    <mrow> <mtable> <mtr> <mtd> <mrow> <mi>S</mi> <mo>=</mo> <mn>1</mn> <mo>-</mo> <mfrac> <mi>K</mi> <mi>V</mi> </mfrac> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msup> <mi>S</mi> <mo>&amp;prime;</mo> </msup> <mo>=</mo> <mn>1</mn> <mo>-</mo> <mfrac> <mi>K</mi> <msup> <mi>V</mi> <mo>&amp;prime;</mo> </msup> </mfrac> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msup> <mi>S</mi> <mo>&amp;prime;</mo> </msup> <mo>=</mo> <mi>S</mi> <mo>-</mo> <mfrac> <mi>K</mi> <msup> <mi>V</mi> <mo>&amp;prime;</mo> </msup> </mfrac> <mo>+</mo> <mfrac> <mi>K</mi> <mi>V</mi> </mfrac> </mrow> </mtd> </mtr> </mtable> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>6</mn> <mo>)</mo> </mrow> </mrow>
    Wherein, S is untreated saturation degree component, and V is non-reinforced luminance component, and S ' is the saturation degree component after processing, and V ' is Luminance component after enhancing, K represent the minimum value of RGB triple channels；
    D3, in practice, strengthening the minimum value of front and rear RGB triple channels can change；Minimum is compensated using Gamma conversion The dynamic change of value；
    So=(S ')γ (7)
    Wherein, SoFor the saturation degree component of output, S ' is the saturation degree component after handling for the first time, and γ is Gamma coefficients.
CN201410736010.1A 2014-12-04 2014-12-04 A kind of local Retinex Enhancement Methods based on HSV color spaces Active CN104537615B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410736010.1A CN104537615B (en) 2014-12-04 2014-12-04 A kind of local Retinex Enhancement Methods based on HSV color spaces

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410736010.1A CN104537615B (en) 2014-12-04 2014-12-04 A kind of local Retinex Enhancement Methods based on HSV color spaces

Publications (2)

Publication Number Publication Date
CN104537615A CN104537615A (en) 2015-04-22
CN104537615B true CN104537615B (en) 2017-12-05

Family

ID=52853134

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410736010.1A Active CN104537615B (en) 2014-12-04 2014-12-04 A kind of local Retinex Enhancement Methods based on HSV color spaces

Country Status (1)

Country Link
CN (1) CN104537615B (en)

Families Citing this family (21)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105303515B (en) * 2015-09-22 2018-06-26 中国科学院上海技术物理研究所 A kind of color misregistration correction method under special illumination condition in closed experimental box
CN106023110B (en) * 2016-05-20 2018-05-29 河海大学 A kind of Hi-Fi image defogging method
CN106251300B (en) * 2016-07-26 2019-04-09 华侨大学 A kind of quick night Misty Image restored method based on Retinex
CN107871303B (en) * 2016-09-26 2020-11-27 北京金山云网络技术有限公司 Image processing method and device
US10692194B2 (en) 2016-09-30 2020-06-23 Huawei Technologies Co., Ltd. Method and terminal for displaying edge of rectangular frame
CN108230233A (en) * 2017-05-16 2018-06-29 北京市商汤科技开发有限公司 Data enhancing, treating method and apparatus, electronic equipment and computer storage media
CN107292860B (en) * 2017-07-26 2020-04-28 武汉鸿瑞达信息技术有限公司 Image processing method and device
CN108205812B (en) * 2017-11-22 2021-12-17 广东工业大学 Method for matching pigment color mixing proportion
CN108122213B (en) * 2017-12-25 2019-02-12 北京航空航天大学 A kind of soft image Enhancement Method based on YCrCb
CN108280800A (en) * 2018-01-29 2018-07-13 重庆邮电大学 A kind of quantum image convolution method
CN108648160B (en) * 2018-05-14 2021-01-15 中国农业大学 Underwater sea cucumber image defogging enhancement method and system
CN109919848B (en) * 2018-07-26 2020-11-17 苏州斯莱斯食品有限公司 Self-checking cabinet of fire extinguishing equipment
CN110246195B (en) * 2018-10-19 2022-05-17 浙江大华技术股份有限公司 Method and device for determining atmospheric light value, electronic equipment and storage medium
CN109978789A (en) * 2019-03-26 2019-07-05 电子科技大学 A kind of image enchancing method based on Retinex algorithm and guiding filtering
CN110321452B (en) * 2019-05-05 2022-08-09 广西师范大学 Image retrieval method based on direction selection mechanism
CN110223253B (en) * 2019-06-10 2023-02-28 江苏科技大学 Defogging method based on image enhancement
CN111311502A (en) * 2019-09-04 2020-06-19 中国科学院合肥物质科学研究院 Method for processing foggy day image by using bidirectional weighted fusion
CN111539905B (en) * 2020-05-15 2023-08-25 北京百度网讯科技有限公司 Method, apparatus, device and storage medium for generating image
CN112950490B (en) * 2021-01-25 2022-07-19 宁波市鄞州区测绘院 Unmanned aerial vehicle remote sensing mapping image enhancement processing method
CN116258653B (en) * 2023-05-16 2023-07-14 深圳市夜行人科技有限公司 Low-light level image enhancement method and system based on deep learning
CN117315053B (en) * 2023-11-28 2024-03-22 国网山东省电力公司淄博供电公司 Visual effect improvement method for old equipment shooting image

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101303766A (en) * 2008-07-09 2008-11-12 北京航空航天大学 Method for rapidly reinforcing color image based on Retinex theory
CN102930512A (en) * 2012-09-25 2013-02-13 哈尔滨工程大学 HSV (Hue, Saturation and Value) color space based underwater image enhancing method by combining with Retinex

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101303766A (en) * 2008-07-09 2008-11-12 北京航空航天大学 Method for rapidly reinforcing color image based on Retinex theory
CN102930512A (en) * 2012-09-25 2013-02-13 哈尔滨工程大学 HSV (Hue, Saturation and Value) color space based underwater image enhancing method by combining with Retinex

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
《HSV色彩空间的Retinex结构光图像增强算法》;秦绪佳等;《计算机辅助设计与图形学学报》;20130430;第25卷(第4期);488-493 *
《基于亮度划分MSR的视觉图像增强》;卢志茂等;《华中科技大学学报（自然科学版）》;20111130;第39卷;99-102 *

Also Published As

Publication number Publication date
CN104537615A (en) 2015-04-22

Similar Documents

Publication Publication Date Title
CN104537615B (en) A kind of local Retinex Enhancement Methods based on HSV color spaces
CN103955905B (en) Based on the single image to the fog method that fast wavelet transform and weighted image merge
CN109754377B (en) Multi-exposure image fusion method
US9569827B2 (en) Image processing apparatus and method, and program
CN101783012B (en) Automatic image defogging method based on dark primary colour
CN108376391A (en) A kind of intelligence infrared image scene Enhancement Method
CN106846263A (en) The image defogging method being immunized based on fusion passage and to sky
CN103218778B (en) The disposal route of a kind of image and video and device
CN102831591B (en) Gaussian filter-based real-time defogging method for single image
CN104537634B (en) The method and system of raindrop influence is removed in dynamic image
WO2016206087A1 (en) Low-illumination image processing method and device
CN103034983B (en) A kind of defogging method capable based on anisotropic filtering
CN103606132A (en) Multiframe digital image denoising method based on space domain and time domain combination filtering
CN107680056A (en) A kind of image processing method and device
CN103268598A (en) Retinex-theory-based low-illumination low-altitude remote sensing image enhancing method
CN104318524A (en) Method, device and system for image enhancement based on YCbCr color space
CN108038833B (en) Image self-adaptive sharpening method for gradient correlation detection and storage medium
CN104346774B (en) Method and apparatus for image enhaucament
CN108305232B (en) A kind of single frames high dynamic range images generation method
CN107358585A (en) Misty Image Enhancement Method based on fractional order differential and dark primary priori
CN103455979A (en) Low illumination level video image enhancement method
CN103886565A (en) Nighttime color image enhancement method based on purpose optimization and histogram equalization
CN104318535B (en) The method, device and mobile terminal of image defogging
CN105931208A (en) Physical model-based low-illuminance image enhancement algorithm
CN103914820A (en) Image haze removal method and system based on image layer enhancement

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