CN101908330A - Method for display equipment with narrow dynamic range to reproduce image with wide dynamic range - Google Patents

Abstract

The invention relates to a method for display equipment with narrow dynamic range to reproduce an image with wide dynamic range. The dynamic range of the image can be compacted by simulating the process of color vision processing information of human eyes and the texture details and colors of scenes in the images with wide dynamic range can be correctly reproduced on the display equipment with narrow dynamic range. The method comprises the following steps of: reading a real luminance value of each pixel RGB (Red, Green and Blue) channel in the image with wide dynamic range; converting the RGB into XYZ value; calculating the adaptation luminance of each pixel; carrying out the chromaticity adaptation adjustment of a light source under each adaptation luminance; simulating a photo-induction cell response mechanism and calculating the relative response value under each adaptation luminance; and converting to an orthogonal opposite space by using PCA (Principal Component Analysis) and simulating an edge phase resistance mechanism of a center-peripheral perceptive cell structure for further processing. The invention can be used for compressing the image with wider dynamic range at high efficiency so as to adapt to the display equipment with narrow dynamic range or paper.

Description

A kind of low-dynamic range display device reproduces the method for high dynamic range images

Technical field

The invention belongs to the reproducing technology field in the Digital Image Processing, be specifically related to the display reproduction method of high dynamic range images on low-dynamic range display device (as display and paper).

Background technology

Image that generally use now, 8 in each passage is referred to as low dynamic range echograms.This type of image can not write down real brightness in the scene, just writes down that visual coding is crossed, limited contrast relation (each passage has only 256 contrasts usually), and the effect of its final scene that writes down must determine according to concrete display device.If the maximum display brightness of a display is 100cd/m2, dynamic range is 100: 1, and so correctly 1cd/m2 can't present the complete effect of shadow of former scene to the effect of shadow between the 100cd/m2 brightness in the displayed scene.Therefore, it is also referred to as device-dependent image.

(High Dynamic Range image then is the relevant image of a kind of scene HDRi) to high dynamic range images, and its dynamic range generally surpasses 1000: 1, and each pixel record is the scene true brightness.Can write down all visions can observed scene (10 ^-6Cd/m ²-10 ⁸Cd/m ²The scene of brightness range).These advantages in view of high dynamic range images, in recent years, industries such as high definition film, recreation special efficacy, Aero-Space, satellite meteorology, medical treatment, printing and traffic finance expect that all high dynamic range images can replace existing low dynamic range echograms, are widely used.But these use the common problem that all can run into, i.e. the problem of correct display reproduction high dynamic range images on existing low-dynamic range display device.

The dynamic range of the display device of general commercialization is all lower now.For example, the about 80cd/m of the maximum display brightness of CRT monitor ², actual dynamic range is not higher than 100: 1 usually; The dynamic range that paper can present is lower, and inferior machine-glazed paper was opened dynamic range about 50: 1, but not the dynamic range of the paper of inferior light has only about in the of 30: 1.If device-dependent low dynamic range echograms is when being presented on these display devices, the change that presents scene that original image writes down is not sufficient to cause the words of visually-perceptible, when high dynamic range images is presented on these show medias, the compression on a large scale of dynamic range, present the relatively large change of brightness of scene, then can make the grain details and the color of performance real scene that the outer variation of vision tolerance all takes place, the true effect of shadow of the high dynamic range scene that high dynamic range images write down has not existed yet.

For for addressing these problems, industry has proposed a kind of method that is referred to as the contrast mapping.Be divided into overall contrast mapping and the mapping of local contrast again according to mapping mode.Overall situation contrast mapping method is when the changing image dynamic range, and each pixel mapping is utilized same mapping curve relation, and local contrast mapping method then can carry out different conversion at the concrete place of each pixel image-region situation.Overall situation contrast mapping method thinks that fundamentally human eye is when observing the image of any brightness range, and human eye only produces a response curve under an adaptability brightness.In fact, human eye can not adapt to a dynamic range and surpass 1000: 1 scene, and for high dynamic range images, therefore the dynamic range of its scene, uses overall contrast mapping then can cause the loss of image texture details much larger than this.What a lot of local contrast mapping methods were then thought each pixel visually influences each other to low-pass filtering character, and the result has produced the halation problem.Human eye vision is when observing real scene, and the existing overall situation is adjusted and also had local the adjustment.For example, the photoresponse ability of the photopigment in the human eye in each photoinduction cell can change along with the variation of brightness, has the influence of eliminating or weakening the light source variation thereby show as the photoinduction cell, and promptly colourity adapts to, and this is the adjustment of overall importance of human eye.And for example, exist the cell (as the upright cell of kinetocyte, outside knee shape, biconjugate etc.) of a series of centers-peripheral perceptual structure in the human visual system, they make that center response message and peripheral response message hinder mutually on the space.These then are the part adjustment relation of vision.In addition, whether the contrast mapping method mostly is in the mapping of luminance channel and handles, seldom take color into account and be affected.

Summary of the invention

The objective of the invention is provides a kind of high dynamic range images display reproduction method of analog color visual processes information process in order to solve existing high dynamic range images display reproduction above shortcomings.

Technical scheme provided by the invention is the method that a kind of low-dynamic range display device reproduces high dynamic range images, it is characterized in that: comprise following steps,

Step 1 reads the intrinsic brilliance value RGB of each pixel in the high dynamic range images earlier, and is converted to tristimulus values XYZ, obtains the image XYZ space;

Step 2, the Y that is provided with among the step 1 gained tristimulus values XYZ is the brightness L of each pixel of high dynamic range images, the adaptability brightness L when utilizing following formula to calculate each pixel of response of photoinduction cell _w(x, y),

When the low-dynamic range display device is paper, for the image of observing on the paper:

L _w(x，y)＝F _G(k，l)*L(x，y)

When the low-dynamic range display device is display, for the image of observing on the display:

L _w(x，y)＝F _G(k，l)*L(x，y)+L _refl

Wherein, (x y) is each pixel of high dynamic range images (x, brightness value y), F to L _G(k is that (k, gauss low frequency filter l) are used for the approximate simulation human eye and are changing the equivalent adaptation brightness that brightness obtains, L for filtering window l) _ReflThe brightness of on the screen of display, reflecting for light source;

Step 3, utilize Metzler matrix with each pixel in the image XYZ space (x, value X y) (x, y), Y (x, y), Z (x, y) forward to each pixel in the response space of cone cell (x, value R y) (x, y), G (x, y), B (x, y):

$\left(\begin{array}{c}R(x,y)\\ G(x,y)\\ B(x,y)\end{array}\right)=M\left(\begin{array}{c}X(x,y)\\ Y(x,y)\\ Z(x,y)\end{array}\right)$ $M=\left(\begin{array}{ccc}0.7982& 0.3389& 20.1371\\ 20.5918& 1.5512& 0.0406\\ 0.0008& 0.0239& 0.9753\end{array}\right)$

Step 4, utilize colourity adaptability model adjust each pixel (x, y) since the light source light spectrum changes in distribution cause change in color, colourity adaptability model is as follows:

R _a(x，y)＝R(x，y)·(D(L _w(x，y))·(R _wr/R _w)+1-D(L _w(x，y)))

G _a(x，y)＝G(x，y)·(D(L _w(x，y))·(G _wr/G _w)+1-D(L _w(x，y)))

B _a(x，y)＝B(x，y)·(D(L _w(x，y))·(B _wr/B _w)+1-D(L _w(x，y)))

Wherein, R _a(x, y), G _a(x, y), B _a(x y) makes cone cell response after the colourity accommodation, R for each pixel of high dynamic range images at the light source light spectrum changes in distribution _w, G _w, B _wFor ringing cell, the original light source cone of high dynamic range images should be worth R _Wr, G _Wr, B _WrBe the light source cone cell response of low-dynamic range display device, D (L _w(x is at each pixel adaptability brightness L when observing each pixel of high dynamic range images y) _w(x, y) adaptedness of following cone cell;

Step 5, simulated light competent cell response mechanism calculates the relative response r under each pixel adaptability brightness _R(x, y), r _G(x, y) and r _B(x, y):

$r_R(x,y)=B_R(x,y)\·\frac{{\left(R_a(x,y)\right)}^{n(x,y)}}{{\left(R_a(x,y)\right)}^{n(x,y)}+{(\mathrm{\σ}\left(L_w(x,y)\right)}^{n(x,y)}}$

$r_G(x,y)=B_G(x,y)\·\frac{{\left(G_a(x,y)\right)}^{n(x,y)}}{{\left(G_a(x,y)\right)}^{n(x,y)}+{(\mathrm{\σ}\left(L_w(x,y)\right)}^{n(x,y)}}$

$r_B(x,y)=B_B(x,y)\·\frac{{\left(R_a(x,y)\right)}^{n(x,y)}}{{\left(B_a(x,y)\right)}^{n(x,y)}+{(\mathrm{\σ}\left(L_w(x,y)\right)}^{n(x,y)}}$

Wherein, B _R(x, y), B _G(x, y), B _B(x, y) the expression cone cell is at each pixel adaptability brightness L _w(x, y) the bleaching index under, σ (L _w(x y) is each pixel adaptability brightness L _w((x y) is index to n, according to each pixel adaptability brightness L for x, adaptability factor y) _w(x y) sets;

Step 6 utilizes principal component analysis (PCA) the photoinduction cell to be rung the value r in space _R(x, y), r _G(x, y) and r _B(x y) is transformed into the value I in quadrature opposition space _{B_w}(x, y), I _{R_g}(x, y), I _{Y_b}(xy), the edge phase resistance mechanism of Simulation Center-peripheral perceptual structure cell is further handled again:

$R_{b\_w}(x,y)=I_{b\_w}(x,y)+\mathrm{\Σ}\underset{i,j\∈\mathrm{\Ω}}{\mathrm{\Σ}}g(I_{b\_w}((x,y)-I_{b\_w}(i,j))f((i-x,j-y)(I_{b\_w}(x,y)-I_{b\_w}(i,j))$

$R_{r\_g}(x,y)=I_{r\_g}(x,y)+\mathrm{\Σ}\underset{i,j\∈\mathrm{\Ω}}{\mathrm{\Σ}}g(I_{r\_g}((x,y)-I_{r\_g}(i,j))f((i-x,j-y)(I_{r\_g}(x,y)-I_{r\_g}(i,j))$

$R_{y\_b}(x,y)=I_{y\_b}(x,y)+\mathrm{\Σ}\underset{i,j\∈\mathrm{\Ω}}{\mathrm{\Σ}}g(I_{y\_b}((x,y)-I_{y\_b}(i,j))f((i-x,j-y)(I_{y\_b}(x,y)-I_{y\_b}(i,j))$

Wherein, g (Δ I) is the contrast influence function of edge phase resistance mechanism, and Δ I represents I _{B_w}(x, y)-I _{B_w}(i, j), f () is the spacial influence function of edge phase resistance mechanism, Ω is that (x y) is the neighborhood scope at center with pixel;

Step 7 utilizes principal component analysis (PCA) the value R in image quadrature opposition space _{B_w}(x, y), R _{R_g}(x, y), R _{Y_b}(x y) goes back to the photoinduction cell and rings the space, utilizes the inverse matrix of the described Metzler matrix of step 3 to go back to XYZ space again, utilizes the inverse matrix of sRGB transition matrix to be transformed into the value of low-dynamic range display device display space RGB at last.

And, in the described step 2, the brightness L that light source reflects on indicator screen _ReflBe calculated as follows:

$L_{\mathrm{refl}}=\frac{k}{\mathrm{\π}}{E}_{\mathrm{amb}}$

Wherein, E _AmbBe the illuminance of light source in the environment of display place, unit is the lux, and k represents the reflection coefficient of indicator screen.

And, in the described step 4, when observing each pixel of high dynamic range images at each luminance adaptation brightness L _w(x, y) the adaptedness D (L of following cone cell _w(x y) is calculated as follows:

D(L _w(x，y))＝F(0.08log ₁₀(1/5L _w(x，y))+0.76)

F is an envirment factor, L _w(x, y)＞10cd/m ²The time, F=1.0; And L _w(x, y)≤10cd/m ²The time, F=0.8.

And in the described step 5, cone cell is at each pixel adaptability brightness L _w(x, y) the bleaching index B under _R(x, y), B _G(x, y), B _B(x y) is calculated as follows:

B _R(x，y)＝10 ⁷/(10 ⁷+L _w(x，y)(R _wr/(R _wr+G _wr+B _wr)))

B _G(x，y)＝10 ⁷/(10 ⁷+L _w(x，y)(G _wr/(R _wr+G _wr+B _wr)))。

B _B(x，y)＝10 ⁷/(10 ⁷+L _w(x，y)(B _wr/(R _wr+G _wr+B _wr)))

And, in the described step 5, each pixel adaptability brightness L _w(x, adaptability factor sigma may (L y) _w(x y) is calculated as follows:

σ(L _w(x，y))＝c·L _w(x，y)/F _l(x，y)

Wherein c is a constant 20; F _l(x y) is calculated as follows:

F _l(x，y)＝0.2T ⁴(x，y)L _w(x，y)+0.1(1-T ⁴(x，y))(L _w(x，y)) ^1/3。

T(x，y)＝1/(L _w(x，y)+1)

And, in the described step 5, index n (x y) is calculated as follows:

$n(x,y)=\frac{a+b{((L_{w\_\max }-L_w(x,y))/(L_{w\_\max }-L_{w\_\min }))}^d}{1+0.5e^{-(L_w(x,y)-L_{w\_\min })}}$

L _{W_max}And L _{W_min}Be respectively high dynamic range images maximum adaptation brightness and minimum adaptability brightness value in vision, a, b and d are the experiment experience value, get 0.75,1.2 and 0.075 respectively.

And in the described step 6, the contrast influence function g (Δ I) of edge phase resistance mechanism is calculated as follows:

$g\left(\mathrm{\ΔI}\right)=1-\frac{1}{\sqrt{2\mathrm{\π}}{\mathrm{\σ}}_I}{e}^{-\frac{{\mathrm{\ΔI}}^2}{2{\mathrm{\σ}}_I^2}}$

Wherein, σ _lBe contrast cutoff in the edge phase resistance mechanism, σ _LGet 0.7.

And in the described step 6, the spacial influence function f () of edge phase resistance mechanism is calculated as follows:

$f(m,n)=\frac{G(m,n,r_{o1})-G(m,n,r_{o2})}{|r_{o1}-r_{o2}|}$

Wherein, G (m, n, r _O1) and G (m, n, r _O2) be the different Gaussian function of radius size, r _O1And r _O2Be respectively that (x y) is the internal diameter and the external diameter at center, r with pixel _O1Get 2% of the image size of handling, r according to experience _O2Get r _O11.6 times.

The invention provides a kind of high dynamic range images display reproduction method of analog color visual processes information process, make the display device of low-dynamic range have regulating power as the human visual system, finally can correctly reproduce the scene in the high dynamic range images, minimize visible loss of detail, and avoid the generation of problems such as halation, and when reproducing grain details, minimize change in color.The present invention adopts said method, the image with high dynamic range that can be correct shows again on the display device or paper of existing low-dynamic range, and method realizes easily, satisfy the requirement of certain actual effect, also can satisfy industries such as high definition film, recreation special efficacy, Aero-Space, satellite meteorology, medical treatment, printing and traffic finance the high dynamic range images demands of applications.

Description of drawings

Fig. 1 is the process flow diagram of the embodiment of the invention.

Embodiment

Below in conjunction with drawings and Examples the present invention is illustrated in further detail.

Referring to Fig. 1, embodiment mainly comprises following steps:

At first determine display device and display environment, as on indoor CRT monitor, showing so indoor illuminance E _AmbBe 200lux, the CRT maximum display brightness is 80cd/m2, and the screen reflection coefficient k is 4%, light source reflecting brightness L on indicator screen in the computing environment _ReflFor the purpose of concrete enforcement reference, provide the further scheme of embodiment, the brightness L that light source reflects on indicator screen _ReflBe calculated as follows:

$L_{\mathrm{refl}}=\frac{k}{\mathrm{\π}}{E}_{\mathrm{amb}}$

Wherein, E _AmbBe the illuminance of light source in the environment of display place, unit is the lux, and k represents the reflection coefficient of indicator screen.

Step 1 reads the intrinsic brilliance value RGB of each pixel in the high dynamic range images earlier, and is converted to tristimulus values XYZ, obtains the image XYZ space.

(unit is cd/m to embodiment according to each passage RGB of the high dynamic range images that reads in ²), convert the XYZ value (cd/m of unit to ²), Y (x, y) dimension brightness L (x, y) dimension are set.Read the intrinsic brilliance value RGB of each pixel in the high dynamic range images, and utilize the sRGB transition matrix to be converted to tristimulus values XYZ, obtain the image XYZ space.The sRGB transition matrix is a prior art, will not give unnecessary details.

Step 2 is calculated the adaptability brightness L of each pixel under the CRT monitor observation condition _w(x, y).

The embodiment implementation is, the Y that is provided with among the step 1 gained tristimulus values XYZ is the brightness L of each pixel of high dynamic range images, the adaptability brightness L when utilizing following formula to calculate each pixel of response of photoinduction cell _w(x, y),

When the low-dynamic range display device is paper, for the image of observing on the paper:

L _w(x，y)＝F _G(k，l)*L(x，y)

When the low-dynamic range display device is display, for the image of observing on the display:

L _w(x，y)＝F _G(k，l)*L(x，y)+L _refl

Wherein, (x y) is each pixel of high dynamic range images (x, brightness value y), F to L _G(k is that (k, gauss low frequency filter l) are used for the approximate simulation human eye and are changing the equivalent adaptation brightness that brightness obtains, L for filtering window l) _ReflThe brightness of on the screen of display, reflecting for light source.

Step 4, utilize colourity adaptability model adjust each pixel (x, y) since the light source light spectrum changes in distribution cause change in color, colourity adaptability model is as follows:

R _a(x，y)＝R(x，y)·(D(L _w(x，y))·(R _wr/R _w)+1-D(L _w(x，y)))

G _a(x，y)＝G(x，y)·(D(L _w(x，y))·(G _wr/G _w)+1-D(L _w(x，y)))

B _a(x，y)＝B(x，y)·(D(L _w(x，y))·(B _wr/B _w)+1-D(L _w(x，y)))

Wherein, R _a(x, y), G _a(x, y), B _a(x y) makes cone cell response after the colourity accommodation, R for each pixel of high dynamic range images at the light source light spectrum changes in distribution _w, G _w, B _wFor ringing cell, the original light source cone of high dynamic range images should be worth R _Wr, G _Wr, B _WrBe the light source cone cell response of low-dynamic range display device, D (L _w(x is at each pixel adaptability brightness L when observing each pixel of high dynamic range images y) _w(x, y) adaptedness of following cone cell.

That is to say, utilize the Metzler matrix high dynamic range images X (x, y), Y (x, y), Z (x, y) value be converted to the cone cell corresponding space response R (x, y), G (x, y), B (x, y); Utilize each pixel adaptability brightness L again _w(x y), calculates the adaptedness D (L of cone cell under suitable respectively answering property brightness _w(x, y)) then, utilizes colourity adaptability model to carry out the colourity accommodation at different adaptability brightness.For ease of concrete implement with reference to for the purpose of, the further scheme of embodiment is provided, when observing each pixel of high dynamic range images at each luminance adaptation brightness L _w(x, y) the adaptedness D (L of following cone cell _w(x y) is calculated as follows:

D(L _w(x，y))＝F(0.08log ₁₀(1/5L _w(x，y))+0.76)

F is an envirment factor, L _w(x, y)＞10cd/m ²The time, F=1.0; And L _w(x, y)≤10cd/m ²The time, F=0.8.

Step 5, simulated light competent cell response mechanism calculates the relative response r under each pixel adaptability brightness _R(x, y), r _G(x, y) and r _B(x, y):

$r_R(x,y)=B_R(x,y)\·\frac{{\left(R_a(x,y)\right)}^{n(x,y)}}{{\left(R_a(x,y)\right)}^{n(x,y)}+{(\mathrm{\σ}\left(L_w(x,y)\right)}^{n(x,y)}}$

$r_G(x,y)=B_G(x,y)\·\frac{{\left(G_a(x,y)\right)}^{n(x,y)}}{{\left(G_a(x,y)\right)}^{n(x,y)}+{(\mathrm{\σ}\left(L_w(x,y)\right)}^{n(x,y)}}$

$r_B(x,y)=B_B(x,y)\·\frac{{\left(R_a(x,y)\right)}^{n(x,y)}}{{\left(B_a(x,y)\right)}^{n(x,y)}+{(\mathrm{\σ}\left(L_w(x,y)\right)}^{n(x,y)}}$

Wherein, B _R(x, y), B _G(x, y), B _B(x, y) the expression cone cell is at each pixel adaptability brightness L _w(x, y) the bleaching index under, σ (L _w(x y) is each pixel adaptability brightness L _w((x y) is index to n, according to each pixel adaptability brightness L for x, adaptability factor y) _w(x y) sets.

For the purpose of concrete enforcement reference, provide the further scheme of embodiment:

(1) cone cell is at each pixel adaptability brightness L _w(x, y) the bleaching index B under _R(x, y), B _G(x, y), B _B(x y) is calculated as follows:

B _R(x，y)＝10 ⁷/(10 ⁷+L _w(x，y)(R _wr/(R _wr+G _wr+B _wr)))

B _G(x，y)＝10 ⁷/(10 ⁷+L _w(x，y)(G _wr/(R _wr+G _wr+B _wr)))。

B _B(x，y)＝10 ⁷/(10 ⁷+L _w(x，y)(B _wr/(R _wr+G _wr+B _wr)))

(2) each pixel adaptability brightness L _w(x, adaptability factor sigma may (L y) _w(x y) is calculated as follows:

σ(L _w(x，y))＝c·L _w(x，y)/F _l(x，y)

Wherein c is a constant 20; F _l(x y) is calculated as follows:

F _l(x，y)＝0.2T ⁴(x，y)L _w(x，y)+0.1(1-T ⁴(x，y))(L _w(x，y)) ^1/3。

T(x，y)＝1/(L _w(x，y)+1)

(3) index n (x y) is calculated as follows:

$n(x,y)=\frac{a+b{((L_{w\_\max }-L_w(x,y))/(L_{w\_\max }-L_{w\_\min }))}^d}{1+0.5e^{-(L_w(x,y)-L_{w\_\min })}}$

L _{W_max}And L _{W_min}Be respectively high dynamic range images maximum adaptation brightness and minimum adaptability brightness value in vision, a, b and d are the experiment experience value, get 0.75,1.2 and 0.075 among the embodiment respectively.

Step 6 utilizes principal component analysis (PCA) the photoinduction cell to be rung the value r in space _R(x, y), r _G(x, y) and r _B(x y) is transformed into the value I in quadrature opposition space _{B_w}(x, y), I _{R_g}(x, y), I _{Y_b}(xy), the edge phase resistance mechanism of Simulation Center-peripheral perceptual structure cell is further handled again:

$R_{b\_w}(x,y)=I_{b\_w}(x,y)+\mathrm{\Σ}\underset{i,j\∈\mathrm{\Ω}}{\mathrm{\Σ}}g(I_{b\_w}((x,y)-I_{b\_w}(i,j))f((i-x,j-y)(I_{b\_w}(x,y)-I_{b\_w}(i,j))$

$R_{r\_g}(x,y)=I_{r\_g}(x,y)+\mathrm{\Σ}\underset{i,j\∈\mathrm{\Ω}}{\mathrm{\Σ}}g(I_{r\_g}((x,y)-I_{r\_g}(i,j))f((i-x,j-y)(I_{r\_g}(x,y)-I_{r\_g}(i,j))$

$R_{y\_b}(x,y)=I_{y\_b}(x,y)+\mathrm{\Σ}\underset{i,j\∈\mathrm{\Ω}}{\mathrm{\Σ}}g(I_{y\_b}((x,y)-I_{y\_b}(i,j))f(i-x,j-y)(I_{y\_b}(x,y)-I_{y\_b}(i,j))$

Wherein, g (Δ I) is the contrast influence function of edge phase resistance mechanism, and Δ I represents poor between two I, as I _{B_w}(x, y)-I _{B_w}(i, j), f () is the spacial influence function of edge phase resistance mechanism, Ω is that (x y) is the neighborhood scope at center with pixel.

Embodiment utilizes PCA (principal component analysis (PCA)) the photoinduction cell to be rung the value r in space earlier _R(x, y), r _G(x, y), r _B(x y) is transformed into the value I in quadrature opposition space _{B_w}(x, y), I _{R_g}(x, y), I _{Y_b}(xy); The edge phase resistance mechanism of Simulation Center-peripheral perceptual structure cell again is to I _{B_w}(x, y), I _{R_g}(x, y), I _{Y_b}(xy) further handle R _{B_w}(x, y), R _{R_g}(x, y), R _{Y_b}(x, y).For the purpose of concrete enforcement reference, provide the further scheme of embodiment:

(1) the contrast influence function g (Δ I) of edge phase resistance mechanism is calculated as follows:

$g\left(\mathrm{\ΔI}\right)=1-\frac{1}{\sqrt{2\mathrm{\π}}{\mathrm{\σ}}_I}{e}^{-\frac{{\mathrm{\ΔI}}^2}{2{\mathrm{\σ}}_I^2}}$

Wherein, σ _lBe contrast cutoff in the edge phase resistance mechanism, σ _LGet 0.7.

(2) in the described step 6, the spacial influence function f () of edge phase resistance mechanism is calculated as follows:

$f(m,n)=\frac{G(m,n,r_{o1})-G(m,n,r_{o2})}{|r_{o1}-r_{o2}|}$

Wherein, G (m, n, r _O1) and G (m, n, r _O2) be the different Gaussian function of radius size, r _O1And r _O2Be respectively that (x y) is the internal diameter and the external diameter at center, r with pixel _O1Get 2% of the image size of handling, r according to experience _O2Get r _O11.6 times.

Step 7 utilizes principal component analysis (PCA) (being PCA) the value R in image quadrature opposition space _{B_w}(x, y), R _{R_g}(x, y), R _{Y_b}(x y) goes back to photoinduction cellular response spatial value, utilizes the inverse matrix of Metzler matrix to forward the value of XYZ space to again, then, utilizes the inverse matrix of sRGB transition matrix that value is forwarded in the low-dynamic range display display space.So just can on the low-dynamic range display, export low dynamic range echograms.The concrete conversion of this step realizes adopting prior art, and the present invention will not give unnecessary details.

Claims (8)

1. a low-dynamic range display device reproduces the method for high dynamic range images, it is characterized in that: comprise following steps, step 1 reads the intrinsic brilliance value RGB of each pixel in the high dynamic range images earlier, and be converted to tristimulus values XYZ, obtain the image XYZ space;

Step 2, the Y that is provided with among the step 1 gained tristimulus values XYZ is the brightness L of each pixel of high dynamic range images, the adaptability brightness L when utilizing following formula to calculate each pixel of response of photoinduction cell _w(x, y),

When the low-dynamic range display device is paper, for the image of observing on the paper:

L _w(x，y)＝F _G(k，l)*L(x，y)

When the low-dynamic range display device is display, for the image of observing on the display:

L _w(x，y)＝F _G(k，l)*L(x，y)+L _refl

Wherein, (x y) is each pixel of high dynamic range images (x, brightness value y), F to L _G(k is that (k, gauss low frequency filter l) are used for the approximate simulation human eye and are changing the equivalent adaptation brightness that brightness obtains, L for filtering window l) _ReflThe brightness of on the screen of display, reflecting for light source;

Step 3, utilize Metzler matrix with each pixel in the image XYZ space (x, value X y) (x, y), Y (x, y), Z (x, y) forward to each pixel in the response space of cone cell (x, value R y) (x, y), G (x, y), B (x, y):

$\left(\begin{array}{c}R(x,y)\\ G(x,y)\\ B(x,y)\end{array}\right)=M\left(\begin{array}{c}X(x,y)\\ Y(x,y)\\ Z(x,y)\end{array}\right)$ $M=\left(\begin{array}{ccc}0.7982& 0.3389& 20.1371\\ 20.5918& 1.5512& 0.0406\\ 0.0008& 0.0239& 0.9753\end{array}\right)$

Step 4, utilize colourity adaptability model adjust each pixel (x, y) since the light source light spectrum changes in distribution cause change in color, colourity adaptability model is as follows:

R _a(x，y)＝R(x，y)·(D(L _w(x，y))·(R _wr/R _w)+1-D(L _w(x，y)))

G _a(x，y)＝G(x，y)·(D(L _w(x，y))·(G _wr/G _w)+1-D(L _w(x，y)))

B _a(x，y)＝B(x，y)·(D(L _w(x，y))·(B _wr/B _w)+1-D(L _w(x，y)))

Wherein, R _a(x, y), G _a(x, y), B _a(x y) makes cone cell response after the colourity accommodation, R for each pixel of high dynamic range images at the light source light spectrum changes in distribution _w, G _w, B _wFor ringing cell, the original light source cone of high dynamic range images should be worth R _Wr, G _Wr, B _WrBe the light source cone cell response of low-dynamic range display device, D (L _w(x is at each pixel adaptability brightness L when observing each pixel of high dynamic range images y) _w(x, y) adaptedness of following cone cell;

Step 5, simulated light competent cell response mechanism calculates the relative response r under each pixel adaptability brightness _R(x, y), r _G(x, y) and r _B(x, y):

$r_R(x,y)=B_R(x,y)\·\frac{{\left(R_a(x,y)\right)}^{n(x,y)}}{{\left(R_a(x,y)\right)}^{n(x,y)}+{(\mathrm{\σ}\left(L_w(x,y)\right)}^{n(x,y)}}$

$r_G(x,y)=B_G(x,y)\·\frac{{\left(G_a(x,y)\right)}^{n(x,y)}}{{\left(G_a(x,y)\right)}^{n(x,y)}+{(\mathrm{\σ}\left(L_w(x,y)\right)}^{n(x,y)}}$

$r_B(x,y)=B_B(x,y)\·\frac{{\left(R_a(x,y)\right)}^{n(x,y)}}{{\left(B_a(x,y)\right)}^{n(x,y)}+{(\mathrm{\σ}\left(L_w(x,y)\right)}^{n(x,y)}}$

Wherein, B _R(x, y), B _G(x, y), B _B(x, y) the expression cone cell is at each pixel adaptability brightness L _w(x, y) the bleaching index under, σ (L _w(x y) is each pixel adaptability brightness L _w((x y) is index to n, according to each pixel adaptability brightness L for x, adaptability factor y) _w(x y) sets;

Step 6 utilizes principal component analysis (PCA) the photoinduction cell to be rung the value r in space _R(x, y), r _G(x, y) and r _B(x y) is transformed into the value I in quadrature opposition space _{B_w}(x, y) I _{R_g}(x, y), I _{Y_b}(xy), the edge phase resistance mechanism of Simulation Center-peripheral perceptual structure cell is further handled again:

$R_{b\_w}(x,y)=I_{b\_w}(x,y)+\mathrm{\Σ}\underset{i,j\∈\mathrm{\Ω}}{\mathrm{\Σ}}g(I_{b\_w}((x,y)-I_{b\_w}(i,j))f((i-x,j-y)(I_{b\_w}(x,y)-I_{b\_w}(i,j))$

$R_{r\_g}(x,y)=I_{r\_g}(x,y)+\mathrm{\Σ}\underset{i,j\∈\mathrm{\Ω}}{\mathrm{\Σ}}g(I_{r\_g}((x,y)-I_{r\_g}(i,j))f((i-x,j-y)(I_{r\_g}(x,y)-I_{r\_g}(i,j))$

$R_{y\_b}(x,y)=I_{y\_b}(x,y)+\mathrm{\Σ}\underset{i,j\∈\mathrm{\Ω}}{\mathrm{\Σ}}g(I_{y\_b}((x,y)-I_{y\_b}(i,j))f(i-x,j-y)(I_{y\_b}(x,y)-I_{y\_b}(i,j))$

Wherein, g (Δ I) is the contrast influence function of edge phase resistance mechanism, and Δ I represents I _{B_w}(x, y)-I _{B_w}(i, j), f () is the spacial influence function of edge phase resistance mechanism, Ω is that (x y) is the neighborhood scope at center with pixel;

Step 7 utilizes principal component analysis (PCA) the value R in image quadrature opposition space _{B_w}(x, y), R _{R_g}(x, y), R _{Y_b}(x y) goes back to the photoinduction cell and rings the space, utilizes the inverse matrix of the described Metzler matrix of step 3 to go back to XYZ space again, utilizes the inverse matrix of sRGB transition matrix to be transformed into the value of low-dynamic range display device display space RGB at last.

2. reproduce the method for high dynamic range images according to the described low-dynamic range display device of claim 1, it is characterized in that: in the described step 2, the brightness L that light source reflects on indicator screen _ReflBe calculated as follows:

$L_{\mathrm{refl}}=\frac{k}{\mathrm{\π}}{E}_{\mathrm{amb}}$

Wherein, E _AmbBe the illuminance of light source in the environment of display place, unit is the lux, and k represents the reflection coefficient of indicator screen.

3. reproduce the method for high dynamic range images according to the described low-dynamic range display device of claim 1, it is characterized in that: in the described step 4, when observing each pixel of high dynamic range images at each luminance adaptation brightness L _w(x, y) the adaptedness D (L of following cone cell _w(x y) is calculated as follows:

D(L _w(x，y))＝F(0.08log ₁₀(1/5L _w(x，y))+0.76)

F is an envirment factor, L _w(x, y)＞10cd/m ²The time, F=1.0; And L _w(x, y)≤10cd/m ²The time, F=0.8.

4. reproduce the method for high dynamic range images according to the described low-dynamic range display device of claim 1, it is characterized in that: in the described step 5, cone cell is at each pixel adaptability brightness L _w(x, y) the bleaching index B under _R(x, y), B _G(x, y), B _B(x y) is calculated as follows:

B _R(x，y)＝10 ⁷/(10 ⁷+L _w(x，y)(R _wr/(R _wr+G _wr+B _wr)))

B _G(x，y)＝10 ⁷/(10 ⁷+L _w(x，y)(G _wr/(R _wr+G _wr+B _wr)))。

B _B(x，y)＝10 ⁷/(10 ⁷+L _w(x，y)(B _wr/(R _wr+G _wr+B _wr)))

5. reproduce the method for high dynamic range images according to the described low-dynamic range display device of claim 1, it is characterized in that: in the described step 5, each pixel adaptability brightness L _w(x, adaptability factor sigma may (L y) _w(x y) is calculated as follows:

σ(L _w(x，y))＝c·L _w(x，y)/F _l(x，y)

Wherein c is a constant 20; F _l(x y) is calculated as follows:

F _l(x，y)＝0.2T ⁴(x，y)L _w(x，y)+0.1(1-T ⁴(x，y))(L _w(x，y)) ^1/3。

T(x，y)＝1/(L _w(x，y)+1)

6. reproduce the method for high dynamic range images according to the described low-dynamic range display device of claim 1, it is characterized in that: in the described step 5, index n (x y) is calculated as follows:

$n(x,y)=\frac{a+b{((L_{w\_\max }-L_w(x,y))/(L_{w\_\max }-L_{w\_\min }))}^d}{1+0.5e^{-(L_w(x,y)-L_{w\_\min })}}$

L _{W_max}And L _{W_min}Be respectively high dynamic range images maximum adaptation brightness and minimum adaptability brightness value in vision, a, b and d are the experiment experience value, get 0.75,1.2 and 0.075 respectively.

7. reproduce the method for high dynamic range images according to the described low-dynamic range display device of claim 1, it is characterized in that: in the described step 6, the contrast influence function g (Δ I) of edge phase resistance mechanism is calculated as follows:

$g\left(\mathrm{\ΔI}\right)=1-\frac{1}{\sqrt{2\mathrm{\π}}{\mathrm{\σ}}_I}{e}^{-\frac{{\mathrm{\ΔI}}^2}{2{\mathrm{\σ}}_I^2}}$

Wherein, σ _lBe contrast cutoff in the edge phase resistance mechanism, σ _LGet 0.7.

8. reproduce the method for high dynamic range images according to the described low-dynamic range display device of claim 1, it is characterized in that: in the described step 6, the spacial influence function f () of edge phase resistance mechanism is calculated as follows:

$f(m,n)=\frac{G(m,n,r_{o1})-G(m,n,r_{o2})}{|r_{o1}-r_{o2}|}$

Wherein, G (m, n, r _O1) and G (m, n, r _O2) be the different Gaussian function of radius size, r _O1And r _O2Be respectively that (x y) is the internal diameter and the external diameter at center, r with pixel _O1Get 2% of the image size of handling, r according to experience _O2Get r _O11.6 times.

Images