WO2013139067A1 - Method and system for carrying out visual stereo perception enhancement on color digital image - Google Patents

Abstract

The present invention relates to the technical field of digital image processing, especially relates to the method and system for carrying out visual stereo perception enhancement on a color digital image. The solution of the present invention is characterized by: extracting the monocular depth information in the image; taking an observation effect of real scenery as an object to simulate and calculate monocular depth information enhancement. As a result, a visual perception brightness value of an image pixel color is simulated and changed into a new value. A value of visual perception saturation and a value of hue are not changed. The visual perception space sense of the image and stereo perception of the scenery are substantially enhanced.

Description

 Method and system for visual stereoscopic perception enhancement of color digital images

Technical field

 The present invention relates to the field of digital image processing, and in particular to a method and system for visual stereoscopic enhancement of color digital images. Background technique

 Since the mid-20th century, important progress has been made in the study of human visual characteristics. Vision is an important function of human understanding of the world. Vision includes "visual" and "sense", which are called visual perception and visual perception. Visual sensation mainly understands the basic properties (such as brightness and color) of people's response to light (visible radiation) from the molecular point of view. It mainly involves physics and chemistry. Visual perception mainly discusses the way people react and react after receiving visual stimuli from the objective world. It studies how to visually form the appearance of people's space in the external world, so it has both psychological factors. Visual perception is a complex process. In many cases, it is difficult to explain the entire perceptual process by relying solely on the retinal images formed by light onto the retina and the known mechanisms of the eye or nervous system. The objective things that people use to visually have a variety of characteristics, the light that they form, the human visual system will produce different forms of response, so the visual perception is divided into brightness perception, color perception, spatial perception and so on. Studies have determined that some properties are related to the physical quantity of the stimulus. For example, the brightness depends on the intensity of the light, and the color depends on the wavelength of the light. However, some characteristics, such as spatial characteristics, have not yet been completely determined by the physical quantity.

 Visual perception characteristics studies have found that when people observe an image, they simultaneously perceive the image as part of a plane and part of a three-dimensional space. This basic psychological phenomenon is called double perceptual reality of images and is a spatial perceptual property. The problem of spatial perception is the problem of deep perception. Humans do not have organs dedicated to sensing distance. The perception of space often depends not only on vision, but also on external conditions called depth cues and internal conditions of the body to help determine the spatial position of the object. These clues include non-visual depth cues, binocular depth cues, and monocular depth cues.

 The depth perception of human space scenes relies mainly on non-visual depth cues and binocular vision. Both eyes Parallax is one of the main causes of stereo perception and depth perception. However, for two-dimensional images that have been imaged by a single-lens optical system, efforts to find true binocular visual depth cues and to enhance display in the image in the form of monocular depth cues are still in progress.

The depth perception of a person's space scene can also be achieved to a certain extent by monocular vision. In monocular vision Some physical conditions of the stimuli themselves can also become clues to the depth of perception and the sense of three-dimensionality under appropriate application conditions. Some of these cues are generated by the characteristics of the imaging optics, such as scene size and distance, linear perspective, and the like. In addition, visual perception due to changes in the light of the imaging environment, such as the distribution of light and shadow, that is, the brightness of the object appears close, the gray object appears far; the color distribution, that is, the blue object appears far, the red and yellow objects appear close; Perspective, that is, the objects in the vicinity appear clear, and the objects in the distance are unclear. If these clues can be quantified and simulated, it should be helpful to enhance the depth (space) perception between the scene and the stereo perception of the object itself. Unfortunately, there has been only a qualitative representation of these perceptions. The corresponding relationship between the physical quantity and the related physical quantity has not been established, and the calculation of the stereoscopic effect of the color digital image cannot be directly guided.

 The study found that increasing the visual perceptual clarity of the image can enhance the visual perception of the image to a certain extent. Some computer programs have this capability, typically such as Adobe Photoshop CS4. In the Smart Sharpen option of the Filter, you can calculate the color difference between each pixel of the image and the adjacent pixels, and use a part of the difference to change the color value of the image to increase the image clarity. the goal of. In this option, you can set the range of adjacent pixels (Radius), the ratio of the color difference (Amount), the way to pre-process the pixels before sharpening (Remove), and also for highlights and shadows. (Shadow) Fade Amount, Tonal Width, and Radius are selected. It is widely believed in the industry that the sharpening function of the program can significantly enhance the sharpness of the image. However, for the purpose of enhancing the visual depth perception of images, the current version of the software lacks corresponding functions. For example, the program currently calculates the color difference between pixels by the RGB three primary color values of the image pixels. This method of enhancing the pixel color will reduce the visual perception saturation and the offset visual perception hue while improving the visual perception brightness of the color. The higher the sharpening intensity, the more obvious the visual perception of the hue deviation caused by this method. Therefore, the current method is also supposed to be and can be improved only for the purpose of increasing image sharpness. In addition, the way in which adjacent pixels are selected is not related to the ambient light characteristics in the sharpened image. Therefore, it is a pity that the change in image sharpness cannot be closely related to depth perception. Also, the softening option before image sharpening is too simple, and the different content in the image cannot be selected for softening and blurring according to the need, thereby the image softening effect limits the sharpness improvement under certain conditions. The extent, and not related to the depth perception of the image, is equally a pity.

Summary of the invention

The inventors of the present invention have proposed and completed the present invention in order to solve the above problems. It is an object of the present invention to provide a method for visual stereoscopic perception enhancement of color digital images.

 It is yet another object of the present invention to provide a system for visual stereoscopic perception enhancement of color digital images. A typical color digital image is a two-dimensional projection of a real scene in the visible light color space at the imaging plane. Since the actual color space of the imaging device is much smaller than the visible color space, the visual perceptual authenticity of the imaged two-dimensional image is also quite different from the perception of the real scene, including spatial depth perception and scene stereo perception. Therefore, the technical solution of the present invention extracts the monocular depth information in the image, and calculates the monocular depth information enhancement with the observation effect of the real scene. As a result, the visually perceived luminance value of the image pixel color is converted into a new value, and the spatial perception of the image perception and the stereoscopic effect of the scene are significantly enhanced.

The method of the present invention are "two color Epic ™" (SecondEposColor ™), referred to as "two-color verse ^{TM" (SECr} ™), primarily for the visual image to enhance a sense of depth perception and stereoscopic scene is adjusted for each pixel spatial visual perception Brightness value. The amount of adjustment depends on various conditions, including: the color perception brightness value of the pixel itself, the projection of the incident light in the imaging plane in the real scene imaging environment, and the visual perception brightness value of the relevant adjacent pixel color in the incident direction of the light, The visually perceived color gray luminance equivalent value of the visually perceived hue of the pixel color of the color grayscale capability and the visually perceived color gray luminance equivalent value of the visual perceptual saturation. According to the method of the present invention, each pixel color of the image maintains the visual perceptual brightness and the visual perceptual hue value while adjusting the visual perceptual brightness value, and the color visual perception brightness difference between the image pixels and the self brightness The ratio is significantly enhanced, that is, the visual perception brightness difference contrast is significantly enhanced, and therefore, the visual perception of the image and the stereoscopic effect of the scene are significantly enhanced. The method and system of the present invention can be used for natural scene imaging such as photography, videography, film, television, video games, or computer generated images, i.e., any image and device that combines colors of red, green, and blue primary colors.

For a digital image that is typically synthesized from three primary colors of red, green, and blue, according to the method of the present invention, after adjusting the visually perceived luminance value of each pixel color, it can be seen by checking the original image to compare the brightness values of the adjusted image. The color visual perception brightness difference of the adjacent pixels is about 50% on average, and the variation range can be controlled by the set amplitude coefficient, and is also controlled by the set allowable range of the absolute value of the brightness change. According to studies, a large sensitivity range of the eye to brightness, of ^{10- 6 - 10 7 cd / m} 2 ( Hom [de] per square meter). But at any point in real life, the ratio of the maximum and minimum brightness perceived by the human eye rarely exceeds 100. The minimum and maximum brightness range in a bright room for 1 100 cd / m ^2, outside of 10- 1000 cd / m ^2, in the evening of 0.01- 1 cd / m ^2. Therefore, according to the method of the present invention, the visually perceived color brightness of the visible light color space and the color image color space are normalized, and the respective brightness parameters of the respective spaces are used as the base, and the respective brightness parameters are linearly transformed to the 0-100 range. Under this condition, according to the technical solution of the present invention, the same color visual perception mode is used for the above two spaces. Brightness, which records the change in color brightness in the dynamic range of the visible light color brightness in the dynamic range of the image color brightness. Assuming that the color space is circular, it is known that the color sRGB color space typically has a projected area of xy chromaticity plane in the CIExyY space that is approximately 35% of the projected area of the spectral trajectory, and is set to the image color space chromatic range diameter, R ₂ For the spectral trace chromatic range diameter, the ratio of the two projected areas:

(Ri/2) ² X π = 0.35 X (R ₂ /2) ² X π, get R corpse 0.59R ₂ , in the case of a circular color space, the dynamic range ratio of the two spatial chromaticities is the same as the dynamic range of the brightness Therefore, it can be assumed that the image color of the sRGB color space has a visually perceived brightness dynamic range that is roughly equivalent to 59% of the visible light color. When the SECr algorithm calculates the average 50% change in the color perception brightness difference of the adjacent pixels of a typical image, it has significantly close to the visible light color space visual perception brightness difference contrast characteristics, so the visual spatial perception of the image and the stereoscopic perception of the scene can be significantly close. Spatial perception of the real environment and stereo perception of real scenes.

 According to the technical solution of the present invention, one of the image monocular depth information of the selected application is the difference in color perceived luminance between pixels. Because the brightness that people perceive from the surface of an object is basically determined by its relationship to the surrounding environment (especially the background). If two objects have similar brightness differences with their respective backgrounds, then they appear to have similar brightness. This choice also facilitates the application of normalized calculations for the dynamic range of the color space luminance.

 As mentioned above, the guiding significance of image dual perception authenticity for color digital image depth perception enhancement calculation is that, if the image is perceived as having three-dimensional authenticity, the image must be reconstructed carefully, and the useful characteristics of the natural field of view need to be simulated as much as possible. The invention adopts the visual perception generated by the change of the imaging environment light described above, mainly based on the visual perceptual brightness of the color digital image, supplemented by the visual perception hue and saturation, and establishes various models to adjust the pixel color as the image monocular depth information. Visually perceived brightness values, for example:

Model 1, the model of the influence of incident ray on the stereo perception of the scene. The illumination light of the real scene imaging environment is one of the important factors that form the three-dimensional sense of the scene and the spatial position perception between the scenes. The change in the angle between the incident ray and the reflected light from the surface of the scene may indicate the surface shape of the illumination position. The brightness of the surface perceived by the observer is related to the angle between the viewing angle and the reflected light. The smaller the angle, the brighter the feeling. The change in brightness of the surface of the object forms a stereoscopic perception of the object, and the incident ray is the primary condition for the change in the brightness of the surface of the object. Model 1 selects the relevant pixels based on the projection of the incident ray on the image plane, and selects the relevant pixels on the condition that the image pixels are adjacent in the incident direction and the neighbors on both sides. Up to three related pixels are selected to represent the parallel characteristics of the incident ray. Calculate the difference in color visual perception brightness of each pixel of the image and the three related pixels, and each difference is weighted by the respective coefficients set by the SECR algorithm to reflect the correlation, that is, the correlation between the surface brightness and the surface shape. The SECr algorithm also sets the actual utilization factor for the sum of the luminance differences to accommodate the need for image enhancement stereo perception of different characteristics. Applying Model 1, the intensity of incident light in the image looks significant Enhancement, the spatial and three-dimensional sense of the scene is also significantly enhanced.

 Model 2, the model of the effect of pixel color visual perception of brightness values on spatial perception. The visual perception produced by the illumination illumination of the illumination environment as described above relates to the brightness and sharpness of the scene, respectively, and the model 2 associates the two, that is, the luminance value of the image pixel is related to the sum of the luminance differences between the pixels calculated by the model 1. Under the condition of illumination light and observation of the same direction of incidence, the result of SECr algorithm is that the higher the pixel color brightness value is, the higher the ratio of the sum of the above brightness difference values is, the more the pixel color brightness value changes, when it is brighter than the relevant pixel Then calculate the increase brightness, which is calculated to reduce the brightness when it is darker than the relevant pixels. Model 2 correlates the change in contrast between luminances of pixels to the brightness of one of the monocular depth cues. The sharpness of the scene with relatively high luminance values in the image is increased, so the closer it appears, the farther it appears, the farther the image is. The visual space perception of the medium scene is significantly enhanced. The correlation is reversed under backlight conditions, with the result that the higher the brightness, the lower the sharpness and therefore the farther it is. Under the condition that the illumination light is approximately perpendicular to the incident, the image pixel having the brightness value set in the SECR algorithm applies the highest proportion of the sum of the brightness difference values, so that the sharpness of the scene is the highest and appears to be the closest, thereby respectively reaching the brightest The ratio of the sum of the pixel-to-luminance differences in the darkest luminance interval is successively decreasing, so the brightest and darkest scenes in the image appear to be relatively far away.

Model 3, Pixel Color Visually perceives the amount of color gray in the brightness effect model. As described above, in the SECr algorithm, L, C, and h of the CIELAB space represent color visual perception of brightness, saturation, and hue angle, respectively. Previous studies have generally suggested that rod cells do not function under bright vision. According to a recent study, rod cells are still active at around 500 cd/m ^{2 ,} and it is believed that with the development of technology, the color vision model will be further improved. According to the technical solution of the present invention, the visually perceived brightness L value of the color image color is composed of two parts of perceived brightness, that is, neutral gray brightness and color gray brightness. The neutral gray brightness range is approximately equivalent to the dark and full-range portions of the bright vision generated by the human eye rod response and the red-green-blue three-color cone response. The color gray luminance range is the high-end portion of the bright vision generated by the red, green, and blue cone response, that is, the portion closer to the spectral light efficiency curve, and for the color digital image, it is the luminance portion above L _Cmaxhl . According to the technical solution of the present invention, the range of the color gray color on the color planes of the respective colors is calculated differently, so in the model 3, the color phase surface color gray brightness range is normalized to calculate the result, as the visual perception color phase visual perception color gray brightness Equivalent value. An equation in which the ratio of the color gray luminance value of the pixel color to the color gray luminance range is multiplied by the color gray luminance equivalent value, as the submodel 1 of the model 3, represents the color gray luminance contribution dominated by the perceived hue in the visual perception brightness of the image color. The share, the color distribution characteristic in the visual perception produced by the above-mentioned imaging environment light change, that is, the blue object appears far, and the qualitative description of the red and yellow objects appearing is converted into a quantitative description.

Recommended by CIE (International Commission on Lumination International Lighting Commission) The standard method converts the CIE xyY parameter of the Munsell new standard system color sample to the CIELAB spatial hue h, brightness L and saturation C values. It can be seen that the higher the sample color saturation value in the same brightness sequence, the brighter the color appears. Their visual perception brightness L values are the same. Therefore, for a color digital image, the technical solution of the present invention calculates a ratio of a visual color saturation saturation value of a pixel color in a color gray luminance range to a maximum saturation degree of a luminance sequence thereof, as a visually perceived color gray luminance equivalent value of visual perceptual saturation. . The equation by which the equivalent value is multiplied by the ratio of the visually perceived color gray luminance value to the color gray luminance range, as the submodel 2 of the model 3, represents the contribution of the color gray luminance dominated by the perceived saturation in the visual perception of the image color. The share, the light and shadow distribution characteristics in the visual perception produced by the above-mentioned imaging environment light changes, that is, the generally bright objects appear close, and the gray objects appear far qualitative descriptions converted into quantitative descriptions.

 Submodel 1 and submodel 2 of model 3 calculate the adjustment amount of the color gray luminance difference of the pixel color, and the neutral gray luminance difference adjustment amount of the pixel color calculated by the model 2 constitutes the image color visual space perception and the scene stereo perception to the real scene. Simulation.

 With the technical solution of the present invention, the visual perception of the image and the stereoscopic effect are simultaneously significantly enhanced. However, the actual partial image content does not need to be significantly clear. For example, for a portrait-based image, the sharpness of the pores, the hair, and the like may have the opposite effect. Therefore, the technical solution of the present invention further includes selectively softening the image-specific content. Algorithm module. The softening calculation uses a 5 X 5 pixel template centered on the target pixel, a Gaussian distribution weighted convolution averaging algorithm, to adjust the visual perception brightness parameter of the target pixel. The adjustment amount is related to various conditions and settings, including: a threshold of visually perceived luminance difference between pixels set in the SECr algorithm; a neighboring pixel of the target pixel, that is, a neighboring pixel of the target pixel in the template (eight in total) has low brightness The pixel convolution threshold at the threshold; the outer circle pixels of the target pixel in the template (16 in total) the pixel convolution threshold whose brightness is lower than the threshold; the pixel color visual perception hue interval of the specific content set in the SECr algorithm; Pixel color visual perception saturation threshold; set the actual use ratio value of the pixel color visual perception brightness adjustment amount calculated above. The Gaussian distribution weighting coefficients used by the template are adjusted based on the threshold setting of the SECr algorithm design based on the typical Gaussian distribution. It is one of the main technical characteristics of the present invention to perform softening convolution calculation using only the visually perceived luminance value of the color, and is one of the reasons that has significant advantages over other softening algorithms.

 A method for visual stereoscopic perception enhancement of a color digital image in accordance with the present invention includes the following steps:

(1) calculating the color boundary of 360 color phase planes of the visual perceptual space of the device displaying the color digital image, extracting the saturation C _maxU and the maximum saturation C _maxhl and its brightness L _{Cmaxhl on the boundary;}

(2) Forward-converted color digital image pixel color R, G, and B values are L of CIELAB space, C and h values, where h is the hue angle, L is the brightness, and C is the saturation;

 (3) Determine the projection position of the incident light in the color digital image under the real scene imaging condition, set the incident light intensity equivalent Al of the target pixel, the numerical range is 0.0-1, typically 0.4-0.6, and the parallel rays of the symmetric position on both sides of the incident light The intensity equivalents Al_l and Al_2 represent the correlation between the brightness of the pixel at the location and the brightness of the target pixel, respectively:

 Al l =(1-Α1)Χ[(90-α)/90]

 Al_2= 1-Α1-Α1_1

 Where ct is the angle between the projection of the incident ray on the image and the perpendicular;

(4) Calculating the luminance difference ΔΙ^ between the target pixel and the related pixel, and the pixel selected in the incident ray of the target pixel and the symmetric direction of both sides thereof are related pixels, and the luminance difference A ₎ between the target pixel and the target pixel is calculated, and typically includes:

 Calculated under the incident light conditions at the upper left:

ALij= ( L _Ai j - LBi-ij-i ) XA1 + (L _Ai ,j - L _B ij-i) X Al—l + (L _Ai ,j - L _Bi-1 ,j) X Al_2 incident directly above Calculated under lighting conditions:

ALi,j= ( L _Ai ,j - L _B ij-i ) XA1 + (L _Ai ,j - L _Bi-1 ,j) X Al_l + (L _Ai ,j - L _Bi+1 ,j) X Al_2 upper right Calculated under square incident light conditions:

ALi,j= ( L _Ai ,j - LBi+ij-i ) XA1 + (L _Ai ,j - L _B ij-i) X Al—l + (L _Ai ,j - L _Bi+1 ,j) X Al_2 Among them, A _1j represents the target pixel in the image, that is, the pixel that regulates the brightness, L _Al , j represents the brightness of the pixel; the central pixel A _1j has a total of 8 pixels around, and the clockwise arrangement from the upper left corner is:

 Bi-i,j-i, Bij_i, Bi+ij_i, Bi+ij, Bi+ij+i, Bij+i, Bi_ij+i, Bi_ij,

 To LB^ indicates the corresponding pixel brightness,

 The incident light conditions are set to 8 kinds, that is, the incident light source is from the upper left, the upper right, the upper right, the right right, the lower right, the lower right, the lower left, and the right left;

(5) Calculating the visual space perception equivalent of the pixel color visual perception brightness _Dj

 (5-1) Under observation of the same direction illumination,

D _LlJ = (L _AlJ /100)

 (5-2) and under the condition of observing reverse illumination,

D _LlJ = (lL _Al0 /100)

 (5-3) Under observation of approximate vertical illumination conditions,

Set L _ZH to the target brightness, which is the brightness value for the highest visual perception resolution, with a range of values from 50 to 95, typically 75-85. If L _Al ,"> L _ZH , D _LlJ = (100-L _AlJ ) / (100- L _ZH )

 If LAij < LZH' Duj = LAIJ / LZH °

(6) Calculate the pixel color visual perception brightness control brightness L _j

 LTIJ = LAi,j+ ALij X Duj KL

 Among them, KL is the set control proportional coefficient, the value range is 0.0-3.0, typically 1.0-2.0.

 (7) Calculating the color gray brightness equivalent of the pixel color visual perception brightness,

( 7-1 ) Calculating the color gray brightness equivalent D _CAIx and the adjusted brightness of the pixel color visual perception hue

L _T 3⁄4j,

The L _{Cmaxhl of the} color plane of the color digital image device is _normalized by the maximum value, and the color gray brightness equivalent D _{CAIx of the} corresponding color phase surface is _obtained .

Pixel color visual perception brightness is greater than L _Cmaxh ^ _调控 calculation brightness:

 LT2i,j = Lli,j+ ((LAij" LcmaxLl)/( 100- LcmaxLl) ^ DcAIx) ^ Κχ2

Where K _T2 is the set control coefficient, the value range is 0.0-3.0, typically 1.0-2.0;

( 7-2 ) Calculate the color gray brightness equivalent D _{BAOx of the} pixel color visual perception saturation and the adjusted brightness

^L T3i,j

Pixel color visual perception brightness is greater than L _Cmaxh ^ _像素 pixel color visual perception saturation C _Al , j color gray brightness equivalent D _BAOx:

DBAOX ⁼ ( CAij C _max Ll )

 Calculate the adjusted brightness:

 LT3i,j = LT2i,j+ ((LAij- LcmaxLl)/( 100- LcmaxLl) ^ DBAOX) ^ Κχ3

Among them, K _T3 is the set control coefficient, the value range is 0.0-3.0, typically 1.0-2.0.

(8) The hue h, the saturation C of the pixel color, and the calculated L _T3⁄4j or L _j value are inversely transformed into the sRGB spaces R, G, and B values and normalized.

 The method according to the invention further comprises the steps of selecting a color visual perceptual hue interval in the image to be softened and setting the relevant control coefficients:

1.1 Set the high-end boundary H _CX and the low-end boundary H _DX of the visual perception hue interval of the color digital image softening content, and the transition zone width K _HCX and K _{HDX of the} outer edges of the interval, H _CX and H _DX value range 0-359 ° , K _HCX and K _HDX values range 0-20, typical 10, the outer edge hue value of the high-end transition zone of the hue interval: HWGX ⁼ HGX+KHGX, low end: HWDX ⁼ HDX-KHDX,

Only the color softening calculation result is retained for the color in the specified hue interval, and the visual perception brightness adjustment amount obtained by calculating the softening in the transition region is smoothly changed from the hue interval boundary to the outer edge calculation adjustment amount. To. ;

1.2 Set color visual perception saturation ratio high threshold C _CAOX and high-end transition zone width coefficient BIcxi, C _GAOX value range 0.40-0.80, typical 0.60-0.70, BI _cx ^ value range 0.00-1, typical 0.10,

Only for the color specified in the above specified hue interval and the saturation ratio value below the threshold C _CAOX , the softening calculation result is obtained, and the color ratio of the saturation ratio in the C _CAOX to C _CAOX + BI _CX1 is calculated by the softening of the visual perception. Adjustment amount, from C _CAOX to C _CAOX + BI _cxl calculation adjustment amount smoothly changes to 0;

 1.3 Set the softening convolution template and related pixel weights,

5 × 5 pixels centered on the softened pixel as a softening convolution template, L _1j represents the center pixel, that is, softened pixels, the following label _1j represents the pixel position on the template, i represents the column, j represents the row, and the pixel weights are respectively set. for:

Li-2, j-23⁄4 2, Li-i, _j-2 is 1, 1^ ₂ is 2, L, Bu is 1, Li+2, j-23⁄4 2,

Li _-2 , j - i is 1, Li - i, j - i is 4, Li, j - i is 4, L, ij - i is 4, Li + 2, j - i3⁄4 1,

L _i-2 j is 2, Li-ij is 4, is 8, L ₁₊ is 4, Li+ ₂ , j is 2.

 Li-2, j+i3⁄4 1, Li-ij+i is 4, Lij+i is 4, LH-ij+i is 4, Li+2, j+i3⁄4 1,

Li-2, j+23⁄4 2, Li-i, j+23⁄4 1, Li, j+ ₂ is 2, LH - is 1, Li+ ₂ , j+2 is 2;

1.4 Set the color perception visual brightness difference threshold L _YU between the relevant pixels on the template.

L _YU 3⁄4 value range 0-100, typically 2-6, the difference between the brightness of the template center pixel and the other pixel is less than L _YU, then the pixel is the effective pixel;

1.5

Setting a convolution threshold S _N of effective pixels with pixels adjacent to the center pixel, a range of values 0 - 32, typically 24-28, gradient 4, setting a convolution threshold S _w of the effective pixels of the pixel spaced apart from the center pixel, value range 0 -24, typical 10-14, gradient 1 or 2, when the effective pixel convolution value of adjacent pixels is larger than Si^^ and the effective pixel convolution value of the spaced pixels is greater than S _w , the template effective pixel convolution average is calculated as The luminance value of the center pixel L _Jpi , _{j ;}

1.6 Setting the Pixel Vision Sensing Brightness Adjustment The practical application of the scale factor B _JX1 ,

B _{JX1 has a} value range of 0.00-1, typically 0.10-0.30,

BJXI

Among them, L _{TON 1} is the brightness value actually applied by the center pixel. A system for performing visual stereoscopic enhancement according to the color digital image of the present invention includes:

 (1) A device color visual perception spatial color phase surface color boundary calculation module displaying a color digital image, including:

 (1-1) The red, green and blue primary color values of the device color space are converted into CIELAB space L, C and h value calculation unit;

 (1-2) device color visual perception spatial color phase surface color boundary extraction unit;

(1-3) Color phase plane color boundary C _maxU smoothing unit.

 (2) Color digital image pixel color RGB mode forward conversion and merge color phase surface and brightness sequence module, including:

 (2-1) A calculation unit that converts the color digital image pixel color RGB three primary color values into the brightness L, saturation C, and hue angle h value of the CIELAB space.

 (2-2) Pixel color phase plane and brightness sequence merging unit, dividing the image color space into 360 reference color phase planes, rounding the h phase h value into the corresponding reference color phase plane, and setting the luminance L range in the color phase plane Divided into 101 reference sequences, rounded up to the corresponding brightness sequence with the brightness L value rounded off.

 (3) A color digital image selective softening calculation module, comprising:

(3-1) Softening the content primary unit, reading the hue interval boundaries H _cx and H _DX of the softening color set in the system and the transition zone widths K _HCX and K _{HDX of the} outer edges of the interval, reading the setting softness The saturation ratio high threshold C _CAOX and the high-end transition width coefficient BI _CX1 of the color are introduced into the softening calculation unit;

(3-2) Softening the content selection and calculation unit, reading the convolution template-related inter-pixel luminance difference threshold L _YU set in the system, reading the effective pixel convolution threshold Si^3⁄4S _w , the softness set in the application system Template and pixel weighting, calculate the softening brightness L _Ipi , j for the qualified pixels and import it into the softening application computing unit;

(3-3) Softening brightness application calculation unit, reading the brightness adjustment amount set in the system, actually applying the proportional coefficient B _m , and calculating the adjusted brightness value L _{TON 1} instead of the original brightness value L _Al , j of the pixel.

 (4) A color digital image pixel color brightness enhancement calculation module, comprising:

 (4-1) setting the pixel equivalent projection unit for the projection of incident light on the image and the brightness of the target pixel in the actual scene imaging environment,

 Set the incident light intensity equivalent of the destination pixel, Al, value 0.4-0.6, parallel light intensity equivalents Al_l and Al_2, symmetrically located on both sides of the incident light.

Al l = (1-Α1) Χ [(90-α) / 90] Al_2 = 1- A1- A1_1

 Where ct is the angle between the projection of the incident ray on the image and the perpendicular;

 (4-2) Difference in luminance between the target pixel and the associated pixel ΔΙ^ calculation unit,

 The incident light conditions are set to 8 kinds, that is, the incident light source is from the upper left, the upper right, the upper right, the right right, the lower right, the lower right, the lower left, and the right left. The difference in brightness between the target pixel and the related pixel ΔΙ^ typically includes:

 Calculated under the incident light conditions at the upper left:

ALij= ( L _Ai j - LBi-ij-i ) X A1 + (L _Ai ,j - L _B ij-i) X Al—l + (L _Ai ,j - L _Bi-1 ,j) X Al_2 directly above Calculated under incident light conditions:

ALi,j= ( L _Ai ,j - L _B ij-i ) X A1 + (L _Ai ,j - L _Bi-1 ,j) X Al_l + (L _Ai ,j - L _Bi+1 ,j) X Al_2 Calculated in the upper right incident light condition:

ALi,j= ( L _Ai ,j - LBi+ij-i ) X A1 + (L _Ai ,j - L _B ij-i) X Al—l + (L _Ai ,j - L _Bi+1 ,j) X Al_2 where A _1j represents the target pixel in the image, that is, the pixel that regulates the brightness, L _Al , j represents the brightness of the pixel; the central pixel A _1j has a total of 8 pixels around, and the clockwise arrangement from the upper left corner is:

 Bi-i,j-i, Bij_i , Bi+ij_i , Bi+ij, Bi+ij+i , Bij+i , Bi_ij+i , Bi_ij,

 To LB^ indicates the corresponding pixel brightness,

 The incident light conditions are set to 8 kinds, that is, the incident light source is from the upper left, the upper right, the upper right, the right right, the lower right, the lower right, the lower left, and the right left;

(4-3) Pixel color visual perception of brightness performance of the scene space perceptual equivalent D _j calculation unit, under the same direction of observation:

D _LlJ = (L _AlJ / 100)

 Under and under the condition of observing reverse illumination:

D _LlJ = (l- L _Al0 / 100)

 Under conditions of observation and vertical illumination:

Set L _ZH to the target brightness, ie the brightness value for the highest visual perception resolution, value 75-85, if L _Al , >> L _ZH , D _LlJ = (100-L _AlJ ) / (100- L _ZH )

 If LAij < LZH' Duj = LAIJ / LZH;

(4-4) pixel color visual perception brightness control brightness L _j calculation unit,

 = LAij+ ALij X Duj X KL

Above, K _L is the set control proportional coefficient, the value is 1.0-2.0;

(4-5) pixel color visual perception hue color gray brightness equivalent D _CAIx and adjusted brightness L _T3⁄4j calculation unit, Read the image device 0° -359° L _{Cmaxhl of each} color phase plane, normalize with the maximum value as the base, and calculate the color gray brightness equivalent D _CAIx of the corresponding phase surface.

Compare the pixel color visual perception brightness L _{Ald is} greater than L _Cmaxhl to calculate the control brightness:

 LT2i,j = Lli,j+ ((LAij- LcmaxLl)/( 100- LcmaxLl) ^ DcAIx) ^ Κχ2

The above K _T2 is the set control coefficient, the value range is 0.0-3.0, typically 1.0-2.0;

(4-6) color gray brightness equivalent D _BAOx and adjusted brightness L _T3⁄4j calculation unit of pixel color visual perception saturation,

Comparing pixel color visual perception brightness L _{Ald is} greater than L _Cmaxhl when calculating pixel color visual perception saturation C _Al , j color gray brightness equivalent D _BAOx:

DBAOx ⁼ ( CAij CmaxLl

 Calculate the adjusted brightness:

 LT3i,j = LT2i,j+ ((LAij- LcmaxLl)/( 100- LcmaxLl) DBAOX) Κτ3

The above K _T3 is the set control coefficient, the value range is 0.0-3.0, typically 1.0-2.0.

(5) A calculation module that inversely transforms the hue h, the saturation C of the pixel color, and the L _T3⁄4j or L _j value after the calculation control into the three primary color values of the sRGB spaces R, G, and B.

 As a preferred technical solution of the present invention, the device color visual perception spatial color phase surface color boundary data library is first completed by the device color visual perception spatial color phase surface color boundary calculation module in the system of the present invention, and the calculation includes:

 (1-1) The red, green and blue primary color values of the device color space are converted into CIELAB spaces L, C, and h values, and the red, green and blue primary color values of the device color space in the system of the present invention are converted into CIELAB space. The L, C, and h value calculation units perform:

 (1-2) Device color visual perception spatial color phase surface color boundary calculation, performed by the device color visual perception spatial color phase surface color boundary extraction unit:

The reference color phase plane is represented by an integer of 0-359, and the h color value of the device color is rounded off and merged into the corresponding reference color phase plane, and the reference luminance sequence is represented by an integer of 0-100, rounded to the corresponding luminance sequence by the L value, and the color phase is extracted. The maximum color saturation value C _maxU of each luminance sequence is calculated as the color boundary of the color phase plane;

(1-3) Color phase surface color boundary smoothing calculation, performed by the color phase surface color boundary ^^^^ smoothing unit: extracting the maximum saturation value C _{maxhl of the} color phase plane having the brightness L _Cmaxhl to the lowest brightness L=0 The color boundary corresponding to the luminance sequence interval is calculated by a standard linear interpolation algorithm to compensate for the maximum saturation C _{maxU of the} original luminance sequence, which is not smoothed or padded. Calculated color boundary C _maxU and The color boundary C _maxU corresponding to the luminance sequence interval of the luminance L _Cmaxhl L=100 indicates the applied color boundary of the color phase plane. DRAWINGS

 Figure 1 shows the non-standard device color visual perception space example color phase surface all merged color coordinates, the middle gray line represents the merged brightness sequence algorithm, the lower gray line represents the linear interpolation to correct the corresponding brightness sequence maximum saturation ^^^^ non-smooth Decrement and make up for missing calculation results, abscissa saturation C, ordinate brightness L.

 Figure 2-1 shows a flow diagram of a method for visual stereoscopic perception enhancement of a color digital image in accordance with a particular embodiment of the present invention, illustrating a device color visual perception spatial color phase surface color boundary calculation process.

 2-2 shows a flow chart of a method for visual stereoscopic perception enhancement of a color digital image in accordance with an embodiment of the present invention, illustrating a color digital image visual perception brightness control calculation flow.

 Figure 3 shows the calculation of the softened convolution template and pixel weights.

 A and b in Fig. 4 show the correlation between the angle of incident light and the brightness of the relevant pixel, respectively.

 Figure 5, a indicates the pixel position, b indicates the relevant pixel when the light is incident on the upper left side, and c indicates the relevant pixel when the light is incident on the upper right side.

 Figure 6-1 Original image, Figure 6-2 Image processed by the method of the present invention.

 Figure 7-1 Original image, Figure 7-2 Image processed by the method of the present invention.

 Figure 8-1 shows a typical system flow for a computer program using the SECr algorithm.

 Figure 8-2 shows the typical system flow of a TV using the SECr algorithm IP.

 Figure 8-3 shows the typical system flow of a TV using the SECr algorithm ASIC.

 Figure 8-4 shows the typical system flow of an electronic device using the SECr algorithm ASIC.

detailed description

 Example 1

 The color digital image visual stereoscopic enhancement process of the present invention is implemented.

 (1) Calculate a device color visual perception spatial color phase surface color boundary data database displaying a color digital image.

The database is completed by the device color visual perception spatial color phase surface color boundary calculation module in the system of the present invention, and the calculation includes: (1-1) Converting the red, green and blue primary color values of the device color space into CIELAB space LC and h values, and transforming the three color values of the device color space red, green and blue into CIELAB space LC and h value calculation units :

 Electronic devices with image display functions are typically sRGB color space and D65 white field, sRGB space RGB three primary color products are available:

y _g , _max =0.60 X _b , _max =0.15

 D65 white field CIEXYZ three 剌 值 value:

X _w =0.950456 Y _w =l Z _w =l .089058

 Calculate the color from the above parameters RGB numerical transformation CIEXYZ three-excited value required 3 X 3 matrix coefficient: Use the RGB value of the three primary colors of the white field of the device and the CIEXYZ triple-excitation value forward transformation formula:

0.9505 Xr,max Xg,max Xb,

 1 .0000 = Yr,max Yg,max Yb,

 1 .0891 Zr,max Zg,max Zb,

 The 3 X 3 matrix coefficients are expressed as the product of the three primary color values and the luminance values:

0.9505—(Χτ,π iax/yr, max) Yr, max ( ^X g^ aax/y"g, max) Yg, max (Xb, nax/yb,max)Yb, max T

1 .0000 = Yr, max Y _g , max Yb, max 1

1 .0891 (Zr, n iax/yr, max) Yr, max (Zg, n aax/y"g, max) Yg, max (Zb, nax/yb, max) Yb, max 1 Calculated by the above equation RGB Luminance values at maximum saturation of the three primary colors:

Υ _Γ , =0.2126 Y _g , =0.7152 Y _b , =0.0722

Apply the above results to calculate the matrix 3 X 3 coefficients:

 Apply the above matrix coefficients to calculate the color of the device RGB three primary colors to the CIEXYZ space X Y and Z three-excited value transformation, and then apply the D65 white field three-excited value to calculate the color XYZ value to the CIELAB space L C and h value transformation.

The typical equipment consists of 8 colors of red, green and blue primary colors, that is, 2 ^{3 x 8} total of 16777216 color scalars, which are successively transformed by the above calculation.

Non-standard equipment calculates the CIEXYZ triacal value when the white field and the maximum saturation of the three primary colors of red, green and blue: The standard spectrophotometer is used to measure the white field triple excitability X _w ' Y _w ' and Z _w according to the conventional specifications. ' Calculate the white field normalization coefficient K _{1 :} Ki = 100 / Y _W '

 Calculate the CIEXYZ triple excitation value of the white field of the equipment:

X _W =X _W , X Ki , Y _W =Y _W , X Ki , Z _W =Z _W , X Ki

Using a standard spectrophotometer, the tri-maximal values of the maximum saturation of the red, green and blue primary colors of the device are measured according to conventional specifications, Χ, Υ and ΖΛ X _g ', Y _g ' and Z _g ', X _b ', Y _b 'and z _b ', and then calculate the CIEXYZ triple stimuli of the three primary colors:

XK , Z _r , _max = Z _r , X Κ

Xg,max— Xg X Kj , Yg,max— Yg X Kj , Zg, _ma x― Zg X Ki

,max= , X Ki , Yb,max ⁼ Yb, X Ki , Zb, _max =Zb, K

 Using the CIEXYZ triple excitability value of the maximum saturation of the three primary colors obtained above, instead of the 3 X 3 matrix coefficient in the above standard method, the calculated white field CIEXYZ triple excitatory value is substituted for the nominal white field of the above device. The CIEXYZ triplet value, using the standard method described above, sequentially converts a total of 16,777,216 colors, G and B values into CIELAB space L, C and h values.

 (1-2) Equipment color visual perception spatial color phase surface color boundary calculation, performed by device color visual perception spatial color phase surface color boundary extraction unit:

The reference color phase plane is represented by an integer of 0-359, and the h color value of the device color hue is rounded off and entered into the corresponding reference color phase plane. The reference luminance sequence is represented by an integer of 0-100, and the color L value is rounded off and merged into the corresponding luminance sequence. The maximum saturation value C _maxU in each luma sequence of the color phase plane is _extracted as the color boundary calculation data of the color phase plane.

(1-3) Color phase surface color boundary smoothing calculation, performed by the color phase surface color boundary ^^^^ smoothing unit: extracting the maximum saturation value C _{maxhl of the} color phase plane having the brightness L _Cmaxhl to the lowest brightness L=0 The color boundary corresponding to the luma sequence interval is calculated by a standard linear interpolation algorithm, and the maximum saturation C _{maxU of the} original luma sequence is non-smoothly decremented or the padding boundary data is missing. The calculated partial color boundary C _maxU together with the color boundary C _maxU corresponding to the luminance sequence interval of the luminance L _Cmaxh ′ L=100 represents the final applied color boundary of the color phase plane. The color phase coordinates and boundary calculations are shown in Figure 1.

The smoothing referred to in the above algorithm, that is, the color luminance sequence is arranged from high to low, and its saturation C _max L^J is in the above sequence and is greater than all the following sequences. If the saturation value is less than the smoothed calculation value, the calculated value is used instead, and the larger than the calculated value is unchanged.

Store the above calculation results as a database. The data is in the order of the first order color phase surface h sequence brightness sequence L Sorted, a total of 36,360 lines. The calculation process of the above step (1) is shown in Figure 2-1.

 (2) The red, green and blue primary color values of the completed color digital image are converted into CIELAB space L, C and h values and merged into the corresponding color phase plane and brightness sequence, and the color digital image pixel color mode is forward-converted and the merged color phase is converted. The face and brightness sequence modules are executed, and the calculations include:

 (2-1) Color digital image Pixel color RGB values are converted to CIELAB space L, C and h values, calculated by the color digital image pixel color RGB values converted to CIELAB space L, C and h values of the calculation unit:

 Use the device nominal image to display the image or the white field and the corresponding parameters of the red, green and blue primary colors embedded in the image itself, and apply the standard algorithm recommended by CIE to convert the color represented by the red, green and blue primary colors of the image pixel into CIEXYZ The value and the CIELAB space L, C and h values, the algorithm program and related parameters are the same as step (1).

 The image displayed on the non-standard device needs to measure the CIEXYZ triple-excitation value of the white level of the computing device and the maximum saturation of the three primary colors of red, green and blue. The algorithm program and related parameters are the same as steps (1).

 (2-2) Pixel color phase plane and luminance sequence merge calculation, performed by pixel color phase plane and luminance sequence merge unit:

 The image color space is divided into 0-359 total 360 reference color phase planes, the h phase h value is rounded off and merged into the corresponding reference color phase plane, and the luminance phase range in the color phase plane is divided into 0-100 total 101 reference sequences, The brightness L value is rounded off and merged into the corresponding brightness sequence, so that the pixel color can be called with the integer hue h and the brightness L, but the original 8-bit floating point data precision of h, L and C is kept unchanged, which is the guarantee for accurately calculating the brightness enhancement. One, also the accuracy requirement to keep the h, L, and C values inversely transformed into RGB values.

 (3) Selectively calculate the color of the relevant content in the image.

 The softening algorithm of the present invention is mainly used for skin color, firstly selecting the skin color interval by color hue, and then distinguishing the skin color by the color saturation value in the interval, determining the effective pixel number and determining the calculation by the target pixel and the correlation pixel brightness difference threshold in the convolution template. Convolution average. Selective softening calculations include:

(3-1) Set the boundary H _cx and H _DX of the visual perception hue interval where the color digital image softening content is located, and the transition zone widths K _HCX and K _{HDX of the} outer edges of the interval, H _{cx is} set to 90°, H _DX is set 340° is suitable for most image softening needs. _Both K _HCX and K _HDX are set to 10 for smooth transitions between softened content and soft content. The outer edge hue value of the high-end transition zone of the hue interval: H _wcx =90° +10° , low end: H _WDX =340° -10°.

Only for h ≤ 90 ° or 11 ≥ 340 ° color phase in the color retention specification softening convolution calculation result L _Jpi , i, Transition zone 90° -100° The color perception of the color phase surface color is calculated by the softening of the visual perception. From 90° to 100°, the adjustment amount is smoothly changed to 0. The 340° to 330° color phase plane also calculates a smooth change in the brightness adjustment amount.

(3-2) Set the color visual perception saturation ratio high threshold C _CAOX and the high-end transition width coefficient BIcxi, C _GAOX to 0.70, which can include most skin color, BI _{CX1 is} set to 0.10, which can basically achieve softening content smoothing transition.

Only for the color specified in the above specified hue interval and the saturation ratio value is below 0.7, the specification softening convolution calculation result L _Ipi ,j, and the visual perception brightness adjustment amount obtained by softening the color with a saturation ratio of 0.7 to 0.8 From 0.7 to 0.8, the adjustment amount is smoothly changed to 0.

 (3-3) Set the softening convolution template and associated pixel weights,

Set 5 × 5 pixels centered on the softened pixel as the softening calculation template, L _1j represents the center pixel, and the following label _1j represents the pixel position on the template, i represents the column, and j represents the row. The pixel weight setting is mainly related to the distance from the center of the template, and the weight is set to:

Li-2, j-23⁄4 2, Li-i, _j-2 is 1, 1^ ₂ is 2, L, Bu is 1, Li+2, j-23⁄4 2,

Li _-2 , j - i is 1, Li - i, j - i is 4, Li, j - i is 4, L, ij - i is 4, Li + 2, j - i3⁄4 1,

L _i-2 j is 2, Li-ij is 4, is 8, L ₁₊ is 4, Li+ ₂ , j is 2.

 Li-2, j+i3⁄4 1, Li-ij+i is 4, Lij+i is 4, LH-ij+i is 4, Li+2, j+i3⁄4 1,

Li-2, j+23⁄4 2, Li-i, j+23⁄4 1, Li, j+ ₂ is 2, LH - is 1, Li+ ₂ , j+2 is 2

 The above is shown in Figure 3.

(3-4) Set the visual perception brightness difference threshold L _YU between related pixels on the template

The invention mainly realizes the goal of softening the skin color of the person in the image by calculating the weighted average of the visual perception brightness of the pixels. The color visual perception brightness difference threshold L _Yl ^ between the pixels is extremely critical, and the setting is 3-5, which is suitable for most images. Requirements. If the difference between the brightness of the template center pixel and the other pixel is less than L _YU , the pixel is _regarded as a valid pixel, which is recorded as one of the basis for calculating the softening. The color of the pixel larger than L _YU is generally the content other than the skin color, and is recorded as the basis for not calculating the softening. one.

(3-5) Set the effective pixel convolution threshold on the template Si^3⁄4S _w and calculate the softening

28, the result is that no more than two of the adjacent pixels 8 or more and the central pixel luminance difference is greater than the threshold L _YU does not calculate the softening.

The convolution threshold S _w of the effective pixels of the pixels spaced apart from the center pixel is set to 14, resulting in spaced pixels The softening is not calculated when there are 5 or more of the 16 and the central pixel luminance difference is greater than the threshold L _YU .

Calculate the sum of the effective pixel weights on the template, that is, the convolution calculation. If the result is greater than Si^3⁄4S _w , calculate the effective pixel average of the convolution as the luminance value L _Ipi ,j after the center pixel is softened. Set the threshold Si^3⁄4S _w to properly distinguish other image content above the skin area from softening, such as eyebrows, eyelashes, hair, and clothing.

 (3-6) Set the scale factor ^ of the practical application of the pixel visual perception brightness adjustment amount.

In order to prevent the skin color from being too smooth after the softening calculation and appear false, the brightness value setting changed by the softening calculation is reserved for the appropriate part, and B _{M is} set to 0.15, which is suitable for most image needs.

LYONGj,i ⁼ Ljpij j + ( LAIJ -Ljpij ) X 0.15

Among them, L _{TON 1} is the brightness value actually applied by the center pixel.

 (4) Calculate image pixel color visual perception brightness enhancement.

 (4-1) Determine the projection position of the incident light in the color digital image under the real scene imaging condition, set the incident light intensity equivalent Al of the target pixel, the value is 0.4-0.6, and the parallel light intensity equivalent of the symmetric position on both sides of the incident light is Al_l And Al_2, indicating the correlation between the pixel brightness of the location and the brightness of the destination pixel:

 Al_l = (1-A1) X [(90-α) / 90]

 Al_2 = 1- Al- AlJ

 Where ct is the angle between the projection of the incident ray on the image and the perpendicular. The above is shown in Figure 4. It is one of the features of the present invention to correlate image color perception brightness enhancement with incident light characteristics of a real scene imaging environment.

(4-2) Calculating the luminance difference ΔΙ^ between the target pixel and the related pixel, the relevant pixel, that is, the adjacent pixel in the incident light direction of the target pixel and the neighboring pixels in the symmetric direction of both sides, and calculating the luminance difference AL _lj between them and the target pixel _; Typical conditions include:

 Calculated under the incident light conditions at the upper left:

ALi,j= ( L _Ai ,j - LBi-ij-i ) X A1 + (L _Ai ,j - L _B ij-i) X Al—l + (L _Ai ,j - L _Bi-1 ,j) X Al_2 is calculated directly under incident light conditions:

ALi,j= ( L _Ai ,j - L _B ij-i ) X A1 + (L _Ai ,j - L _Bi-1 ,j) X Al—l + (L _Ai ,j - L _Bi+1 ,j) X Al_2 is calculated under incident light conditions at the upper right:

ALi,j= ( L _Ai ,j - LBi+ij-i ) X A1 + (L _Ai ,j - L _B ij-i) X Al—l + (L _Ai ,j - L _Bi+1 ,j) X Above Al_2, A _{1 j} represents the target pixel in the image, that is, the pixel that regulates the brightness, and L _Al , j represents the brightness of the pixel; Up to 8 pixels of B ₁₊₁ and _J+1 are related pixels, LBW^ to L _BL+1 , _J+1 indicates the brightness of the corresponding pixel, and 8 kinds of incident light conditions are set, that is, the incident light source is from the upper left, the upper side, Top right, right right, bottom right, down, left bottom, and right left. The above is shown in Figure 5.

(4-3) Calculating the visual space perception equivalent of the pixel color visual perception brightness DL _X

 (4-3-1) and observation of the same direction of illumination:

D _LlJ = (L _ALJ / 100)

 (4-3-2) and under observation of reverse illumination conditions:

D _LlJ = (l- L _AL0 / 100)

 (4-3-3) and observed under approximate vertical lighting conditions:

Set L _ZH to the target brightness, which is the brightness value for the highest visual perception resolution, with a value of 75-85, suitable for most image needs.

If L _AL ,"> L _ZH , D _LlJ = (100-L _ALJ ) / (100- L _ZH )

 If LAij < LzH' DLij = LAij /LZH

 Under the condition of accurately setting the incident light source, the image color brightness contrast adjustment amount is related to the color brightness, that is, the image sharpness improvement is related to the neutral gray brightness value of the color, and the depth of the scene in which the scene is different in the image is realized. Differently and to enhance the stereoscopic effect of the scene, setting the adjustment of the correlation can change the depth perception of the image space and the stereoscopic perception of the scene to a certain extent. This is also one of the features of the present invention.

(4-4) Calculate the pixel color visual perception brightness control brightness L _TX ,

 Liij = LAi,j+ ALij X DLij X KL

 Among them, KL is the set proportional coefficient, the value is 1.0-2.0.

 (4-5) Calculating the color gray brightness equivalent of the pixel color visual perception brightness

(4-5-1) Calculate the color gray brightness equivalent D _CAIx and the adjusted brightness L _T 3⁄4j of the pixel color visual perception hue,

L _CMAXHL which _{displays the} phase planes of the color digital image devices from 0° to 359° is normalized with the maximum value as the base, and the color gray luminance equivalent D _{CAIx of the} phase planes of the respective colors is obtained. For the sRGB color space D65 white field color, 103 ° yellow D _CAIx maximum is 1, 306 ° blue D _CAIx minimum is 0.3331.

Pixel color visual perception brightness L _Al , j is greater than L _Cmaxh ^ _调控 calculation brightness:

 LT2i,j = Lli,j+ ((LAij" LcmaxLl)/( 100- LcmaxLl) ^ DcAIx) ^ Κχ2

Among them, K _T2 is the set control coefficient, the value is 1.0-2.0.

(4-5-2) Calculate the color gray brightness equivalent D _BAOx and the adjusted brightness of the pixel color visual perception saturation ^L T3i,j '

Pixel color visual perception brightness L _{Ald is} greater than L _Cmaxh ^ _像素 pixel color visual perception saturation C _Al , j color gray brightness equivalent D _BAOx:

DBAOx ⁼ ( CAij CmaxLl )

 Calculate the adjusted brightness:

 LT3i,j = LT2i,j+ ((LAij- LcmaxLl)/( 100- LcmaxLl) DBAOX) Κτ3

Among them, K _T3 is a regulation coefficient with a value range of 0.0-3.0, typically 1.0-2.0.

 On the basis of the above enhancement of the color neutral gray brightness, the image sharpness enhancement is further correlated with the color gray light brightness value of the color, thereby enhancing the spatial depth of the scene in which the scene is different in the image and enhancing the scene stereo. The effect of feeling. This is also one of the features of the present invention.

(5) The values of h, C, and L _T3⁄4j or L _{j of the} pixel color are inversely transformed into sRGB spaces R, G, and B values. According to a particular embodiment of the invention, the image pixels are color saturated. The hue h, and the luminance L _T3⁄4j or L _j values are calculated as the specifications 1, G and B values. According to a preferred technical solution of the present invention, an image pixel color mode inverse transform and normalization module in the system of the present invention is invoked, and the calculation includes:

 Using the standard method recommended by the CIE, the image pixel color is forward transformed to the unchanging saturation. The hue angle h value and the brightness obtained by the calculation enhancement 1^ or 1^," are calculated as the three primary colors of the device, red, green and blue. This algorithm is the inverse of the above step (2), calculated by the pixel color CIELAB parameter

The white field required for CIEXYZ triple excitatory value CIEXYZ triple excitatory value is the same as the forward calculation, by pixel color

CIEXYZ triacal value calculation RGB value required 3 X 3 matrix coefficient is obtained by inverting the matrix 3 X 3 system number used in the above step (2):

 - 3. 2406 - 0. 9689 0. 0557 - - 1. 5372 1. 8758 - 0. 2040

 - 0. 4986 0. 0415 1. 0570 The calculated R, G and B values are rounded, respectively, and the value greater than 255 is normalized to 255, and the value less than 0 is normalized to 0. Embodiment 2 A typical system using the SECr algorithm of the present invention

 (1) Typical system flow of computer program using SECr algorithm

The computer hard disk HD is used as a typical carrier for submitting the SECR algorithm, and the same function carrier also includes a CD, a DVD, a USB disk, etc., and is authorized to call the SECR algorithm by the network. The SECr algorithm is called programmatically by the computer CPU + GPU and runs in RAM. Color digital images are stored on your computer's hard drive Called by the SECR algorithm program, processed by the SECR algorithm and then stored back to the hard disk. The image data can be copied to a carrier such as a CD, a DVD, a USB flash drive, or another hard disk, or can be transmitted to a specified location through a network.

 The SECr algorithm program can process single frame images and frame sequence images. The single-frame image format can be an uncompressed format such as .tif or .bmp, or a compression format such as .jpg. The frame sequence image format can be general. MOV, .AVI, etc., and can also use the dedicated I/O to process related format files. Watching the SECr algorithm in real time The number of displays that transform image effects can be configured as needed. The system is shown in Figure 8-1.

 (2) Typical system flow using the SECr algorithm IP

 Using the TV main chip as the typical application of the SECr algorithm IP, the R, G and B color gradation lookup tables used by the gamma correction module in IP can be adjusted according to the special gamma settings in the main chip of the TV. The system is shown in Figure 8-2.

 (3) Typical system flow of TV using SECr algorithm ASIC

 Using a TV as a typical application of the SECr algorithm ASIC, set the I/O matching video image to obtain the video image color RGB data. The system is shown in Figure 8-3.

 (4) Typical system flow of electronic equipment using SECr algorithm ASIC

 Applications using the FECr algorithm ASIC also include notebook computers, tablet computers, mobile phones, game consoles, LCD monitors, computer graphics cards, etc., as shown in Figure 8-4.

 The above specific embodiments are only used to illustrate the technical solutions of the present invention and are not intended to be limiting, and the present invention will be described in detail with reference to the above embodiments, and those skilled in the art should understand that the technical solutions of the present invention are modified or equivalently replaced. The spirit and scope of the technical solutions of the present invention are intended to be within the scope of the appended claims.

Claims

Rights request

 A method for performing visual stereoscopic enhancement on a color digital image, the method comprising the steps of:

(1) calculating the color boundary of 360 color phase planes of the visual perceptual space of the device displaying the color digital image, extracting the saturation C _maxU and the maximum saturation C _{maxhl on the boundary} and its brightness L _Cmaxhl;

 (2) Forward-converted color digital image pixel color R, G, and B values are the L, C, and h values of CIELAB space, where h is the hue angle, L is the brightness, and C is the saturation;

 (3) Determine the projection position of the incident light in the color digital image under the real scene imaging condition, set the incident light intensity equivalent Al of the target pixel, the numerical range is 0.0-1, typically 0.4-0.6, and the parallel rays of the symmetric position on both sides of the incident light The intensity equivalents are Al_l and Al_2, which indicate the correlation between the pixel brightness at that location and the brightness of the destination pixel:

 Al l =(1-Α1)Χ[(90-α)/90]

 Al_2= 1-Al-AlJ

 Where ct is the angle between the projection of the incident ray on the image and the perpendicular;

 (4) Calculate the luminance difference ΔΙ^ between the target pixel and the related pixel, and the relevant pixel, that is, the adjacent pixel in the direction of the light incident and the neighboring pixels in the symmetric direction of both sides, calculate the luminance difference Δ between them and the target pixel.

 Calculated under the incident light conditions at the upper left:

AL _i j=(L _A ij - LBi-ij-i) XA1 + (L _Ai ,j - L _B ij-i) X A1—1 + (L _Ai ,j - L _Bi-1 ,j) X Al_2 Positive Calculated under the incident light above:

AL _i j=(L _A ij - L _B ij-i) XA1 + (L _Ai ,j - L _Bi-1 ,j) X Al_l + (L _Ai ,j - L _Bi+ ij) Al_2 Under the right upper incident light condition Calculation:

AL _i j=(L _A ij - LBi+ij-i) XA1 + (L _A ij - L _B ij-i) X Al_l + (L _Ai ,j - L _Bi+ ij) Al_2 where A _1j represents the purpose in the image Pixels, that is, pixels that regulate brightness, L _Al , j represent the brightness of the pixel; a total of 8 pixels around the center pixel A _1j , clockwise from the upper left corner are:

 Bi-i,j-i, Bij_i, Bi+ij_i, Bi+ij, Bi+ij+i, Bij+i, Bi_ij+i, Bi_ij,

 To LB^ indicates the corresponding pixel brightness,

 There are 8 kinds of incident light conditions, which are related to the above pixel positions, that is, the incident light source is from the upper left, the upper right, the upper right, the right right, the lower right, the lower right, the lower left and the right left;

(5) Calculating the visual space perception equivalent of the pixel color visual perception brightness _Dj

(5-1) Under the same illumination conditions as observation: D _LlJ = (L _AlJ / 100)

 (5-2) Under observation of reverse illumination conditions:

D _LlJ = (l- L _Al0 / 100)

 (5-3) Under observation of approximate vertical illumination conditions:

Set L _ZH to the target brightness, which is the pixel brightness value for the highest visual perception resolution, with a range of 50-95, typically 75-85.

If L _Al ,"> L _ZH , D _LlJ = (100-L _AlJ ) / (100- L _ZH )

 If LAij < LZH' Duj = LAij / LZH;

(6) Calculate the brightness of the pixel color visual perception brightness control L _Tl ,j,

 LTIJ = LAi,j+ ALij X Du,j KL

 Where KL is the set regulation factor, the value range is 0.0-3.0, typically 1.0-2.0;

 (7) Calculate the color gray brightness equivalent of the pixel color visual perception brightness,

( 7-1 ) Calculating the color gray brightness equivalent D _CAIx and the adjusted brightness of the pixel color visual perception hue

L _T 3⁄4j,

The L _{Cmaxhl of the} color surface of the color digital image device is _normalized by the maximum value, and the color gray brightness equivalent D _{CAIx of each} color phase surface is _obtained .

Pixel color visual perception brightness L _Al , j is greater than L _Cmaxh ” Calculated brightness after adjustment:

 LT2i,j = Lli,j+ ((LAij" LcmaxLl)/( 100- LcmaxLl) ^ DcAIx) ^ Κχ2

Where K _T2 is the set control coefficient, the value range is 0.0-3.0, typically 1.0-2.0,

(7-2) Calculate the color gray brightness equivalent D _BAOx and the adjusted brightness of the pixel color visual perception saturation

^L T3i,j'

Pixel color visual perception brightness L _{Ald is} greater than L _Cmaxh ^ _像素 pixel color visual perception saturation C _Al , j color gray brightness equivalent D _BAOx:

DBAOx ⁼ ( CAij CmaxLl )

 Calculate the adjusted brightness:

 LT3i,j = LT2i,j+ ((LAij- LcmaxLl)/( 100- LcmaxLl) ^ DBAOX) ^ Κχ3

Where K _T3 is the set control coefficient, the value range is 0.0-3.0, typically 1.0-2.0;

(8) The hue h, the saturation C, and the calculated adjusted luminance L _T3⁄4j or L _j values of the pixel color are inversely transformed into the sRGB spaces R, G, and B values and normalized.

2. The method according to claim 1, wherein the device color visually perceives the spatial color phase surface color boundary and the upper saturation value thereof, and is obtained by the following method:

2.1 For equipment that displays color digital images using the three primary colors of red, green and blue, use the nominal white field and red, green and blue primary color parameters, and use the standard method recommended by CIE to synthesize the three primary colors of red, green and blue. All colors, converted to CIEXYZ triple stimuli and CIELAB space LC and h values, including:

 Calculate the maximum saturation of RGB three primary colors with sRGB space RGB three primary colors and D65 white field CIEXYZ triple value:

Υ _Γ , =0.2126 Y _g , =0.7152 Y _b , =0.0722

Apply the above results to calculate the 3 X 3 matrix coefficients required to convert the color RGB values into CIEXYZ triple excitons:

 2.2 Non-standard equipment shall calculate the CIEXYZ triple-excitation value when the equipment is white field and the maximum saturation of the three primary colors of red, green and blue:

2.2.1 Using a standard spectrophotometer, measure the white field triple excitability values X _w ' Y _w ' and Z _w ' according to the conventional specifications, and calculate the white field normalization coefficient K _{1 :}

Ki = 100 / Y _W '

 Calculate the CIEXYZ triple excitation value of the white field of the equipment:

2.2.2 Using a standard spectrophotometer, measure the three stimuli values of the maximum saturation of the red, green and blue primary colors according to the conventional specifications, Χ Υ and ΖΛ X _g ' Y _g ' and Z _g ' X _b ' Y _b 'and z _b ' then calculate the CIE XYZ triple stimuli of the three primary colors separately:

Xg,max— Xg X Kj Yg,max— Yg X Kj Zg _ma x― Zg X Ki

Max= X Ki Yb,max ⁼ Yb XK Zb, _max =Zb XK

 2.3 Using the three primary color CIEXYZ triple enthalpy values obtained above, instead of the 3 X 3 matrix coefficients in the standard method described in step 2.1, convert the color RGB values into CIEXYZ triple stimuli values to calculate the white field CIEXYZ triple stimuli The value replaces the device nominal white field CIEXYZ triple excitability value in step 2.1, and calculates the color CIEXYZ triple excitability value as CIELAB space LC and h value;

2.4 Calculate the total color LC and h values of the device space, round off the h phase h value Merging into the corresponding reference color phase plane, and then rounding up to the corresponding brightness sequence with the brightness L value;

2.5 The maximum saturation C _maxU value of all luminance sequences in the color phase plane indicates the color boundary of the color phase plane, and the brightness from the maximum saturation value C _{maxhl of the} color phase plane to the brightness L _Cmaxhl to the lowest brightness L=0 The color boundary corresponding to the sequence interval is corrected by the standard linear interpolation method. The non-smooth diminishing or padding boundary data of ^^ is missing, and the calculated C _{maxU is taken} as the applied color boundary.

 3. The method according to claim 1, wherein the color digital image pixel color 1 G and B values are converted into L C and h parameters in the CIELAB space, obtained by:

3.1 Using the equipment or image nominal white field and G and B three primary color parameters, using the standard method recommended by CIE, the image pixel RGB three primary color values are calculated as CIEXYZ triple excitation value and CIELAB space LC and h value, including:

 Calculate the brightness value of RGB three primary colors with maximum saturation in sRGB space RGB three primary colors and D65 white field CIEXYZ triple value:

Υ _Γ , =0.2126 Y _g , =0.7152 Y _b , =0.0722

Apply the above results to calculate the 3 X 3 matrix coefficients required to convert the color RGB values into CIEXYZ triple excitons:

 3.2 Displaying images on non-standard equipment It is necessary to calculate the CIEXYZ triple-excitation value when the white field of the equipment and the maximum saturation of the three primary colors of R G and B are:

3.2.1 Using a standard spectrophotometer, measure the white field triple excitability values X _w ' Y _w ' and z _w ' according to the conventional specifications, and calculate the white field normalization coefficient κ _{1 :}

Ki = 100 / Y _W '

 Calculate the CIEXYZ triple excitation value of the white field of the equipment:

3.2.2 Using a standard spectrophotometer, measure the three stimuli values of the maximum saturation of the RGB three primary colors of the device according to the conventional specifications: XY and ΖΛ X _g ' Y _g ' and Z _g ' X _b ' Y _b ' and Z _b 'Then calculate the CIEXYZ triple stimuli of the three primary colors separately:

Xg,max— Xg X Kj Yg,max— Yg X Kj Zg _ma x― Zg X Ki

X Kl Yb,max ⁼ Yb ' X Kl Zb, _max =Zb ' X Ki ,

 3.3 Using the CIEXYZ triple excitability value of the maximum saturation of the three primary colors obtained above, replacing the 3 X 3 matrix coefficient in the standard method described in step 3.1, using the calculated white field CIEXYZ triple excitatory value instead of the device standard described in step 3.1 The white field CIEXYZ triple value is calculated, and the color RGB value is calculated step by step into the CIELAB space L, C and h values.

 4. The method according to claim 1, wherein the method comprises the steps of selecting a color visual perception hue interval of the content to be softened in the image and setting the relevant control coefficient:

4.1 Set the high-end boundary H _cx and the low-end boundary H _DX of the visual perception hue interval of the color digital image softening content, and the transition zone widths K _HCX and K _{HDX of the} outer edges of the interval, H _cx and H _DX value range 0-359 ° , K _HCX and K _HDX values range 0-20, typical 10, the outer edge hue value of the high-end transition zone of the hue interval: HWGX ⁼ HGX+KHGX, low end: HWDX ⁼ HDX-KHDX,

 Only the color softening calculation result is retained for the color in the specified hue interval, and the visual perception brightness adjustment amount obtained by the softening of the color in the transition region is smoothly changed from the hue interval boundary to the outer edge calculation adjustment amount. ;

4.2 Set color visual perception saturation ratio high limit threshold C _CAOX and high-end transition zone width coefficient BIcxi, C _GAOX value range 0.40-0.80, typical 0.60-0.70, BI _cx ^ value range 0.00-1, typical 0.10,

Only for the color specified in the above specified hue interval and the saturation ratio value below the threshold C _CAOX , the softening calculation result is obtained, and the color ratio of the saturation ratio in the C _CAOX to C _CAOX + BI _CX1 is calculated by the softening of the visual perception. Adjustment amount, from C _CAOX to C _CAOX + BI _cxl calculation adjustment amount smoothly changes to 0;

 4.3 Set the softening convolution template and related pixel weights,

5 × 5 pixels centered on the softened pixel as a softening convolution template, L _1j represents the center pixel, that is, softened pixels, the following label _1j represents the pixel position on the template, i represents the column, j represents the row, and the pixel weights are respectively set. for:

Li-2, j-23⁄4 2, Li-i, _j-2 is 1, 1^ ₂ is 2, L, H, j-23⁄4 1, Li+ ₂ , j _-2 is 2

Li _-2 , j - i is 1, Li - i, j - i is 4, Li, j - i is 4, L, hij - i is 4, Li + 2, j - i3⁄4 1,

L _i-2 j is 2, Li-i j is 4, is 8, L ₁₊ is 4, Li+ ₂ , j is 2

 1, Li-ij+i is 4, Lij+i is 4, LH-ij+i is 4, Li+2, j+i3⁄4 1,

Li-2, j+23⁄4 2 , Li-i, j+23⁄4 1, Li, j+ ₂ is 2, LH - is 1, Li+ ₂ , j+2 is 2;

4.4 Set the color perception visual brightness difference threshold L between related pixels on the template L _YU 3⁄4 value range 0-100, typically 2-6, the difference between the brightness of the template center pixel and the other pixel is less than L _YU, then the pixel is the effective pixel;

4.5

Setting a convolution threshold S _N of effective pixels with pixels adjacent to the center pixel, a range of values 0 - 32, typically 24-28, gradient 4, setting a convolution threshold S _w of the effective pixels of the pixel spaced apart from the center pixel, value range 0 -24, typical 10-14, gradient 1 or 2, when the effective pixel convolution value of adjacent pixels is larger than Si^^ and the effective pixel convolution value of the spaced pixels is greater than S _w , the template effective pixel convolution average is calculated as The luminance value of the center pixel L _Jpi , _{j ;}

4.6 Set the pixel visual perception brightness adjustment amount of the actual application of the scale factor B _JX1 ,

B _{JX1 has a} value range of 0.00-1, typically 0.10-0.30,

BJXI

Among them, L _{TON 1} is the brightness value actually applied by the center pixel.

 5. A system for visual stereoscopic perception enhancement of a color digital image, wherein the system comprises:

 (1) A device color visual perception spatial color phase surface color boundary calculation module displaying a color digital image, including:

 (1-1) The red, green and blue primary color values of the device color space are converted into CIELAB space L, C and h value calculation unit;

(1-2) device color visual perception spatial color phase surface color boundary extraction unit, rounding up the device color hue h value into the corresponding reference color phase surface, rounding up the brightness L value and entering the corresponding brightness sequence, extracting the color phase surface brightness The maximum saturation value C _{maxU of the} sequence color, the basic data is calculated as the color boundary of the color phase plane;

(1-3) Color phase surface color boundary ^^^^ Smoothing unit, selecting the maximum saturation value C _{maxhl of the} color phase surface, the color boundary C _maxU corresponding to the brightness sequence interval from the brightness L _Cmaxhl to the lowest brightness L=0 Calculate the smoothing boundary by the standard linear interpolation algorithm, make up for the C _maxU non-smooth decreasing or the missing boundary data missing, calculate the color boundary and the color boundary corresponding to the brightness sequence interval of the brightness L _Cmaxhl L=100, indicating the color phase surface Apply color boundary C _{maxU ;}

 (2) Color digital image pixel color RGB mode forward conversion and merge color phase surface and brightness sequence module, including:

(2-1) Convert color digital image pixel color RGB values to L, C, and h values of CIELAB space a calculation unit, where h is a hue angle, L is a brightness, and C is a saturation;

 (2-2) Pixel color phase plane and brightness sequence merging unit, dividing the image color space into 360 reference color phase planes, rounding the h phase h value into the corresponding reference color phase plane, and setting the luminance L range in the color phase plane Divided into 101 reference sequences, rounded up to the corresponding brightness sequence with the brightness L value rounded off;

 (3) A color digital image selective softening calculation module, comprising:

(3-1) Softening the content primary unit, reading the hue interval boundaries H _cx and H _DX of the softening color set in the system and the transition zone widths K _HCX and K _{HDX of the} outer edges of the interval, reading the setting softness The saturation ratio high threshold C _CAOX and the high-end transition width coefficient BI _CX1 of the color are introduced into the softening calculation unit;

(3-2) Softening the content selection and calculation unit, reading the convolution template set in the system and the correlation luminance difference threshold L _YU between the pixels, reading the effective pixel convolution threshold Si^3⁄4S _w , set in the application system Softening the template and pixel weighting, calculating the softening brightness L _Ipi , j for the eligible pixels and importing into the softening application computing unit;

(3-3) Softening brightness application calculation unit, reading the brightness adjustment amount set in the system, actually applying the proportional coefficient B _m , and calculating the adjusted brightness value L _{TON 1} instead of the original brightness value L _Al , _j of the pixel _;

 (4) A color digital image pixel color brightness enhancement calculation module, comprising:

 (4-1) setting the projection angle of the incident light on the image in the real scene imaging environment and the pixel equivalent equivalent calculation unit related to the target pixel brightness,

 Set the incident light intensity equivalent of the destination pixel Al, the value is 0.4-0.6, and the parallel light intensity equivalents of the symmetric symmetry of the incident ray are Al_l and Al_2:

 Al l = (1-Α1) Χ [(90-α) / 90]

 Al_2 = 1- Al- AlJ

 Above, ct is the angle between the projection of the incident ray on the image and the perpendicular;

 (4-2) Difference in luminance between the target pixel and the associated pixel ΔΙ^ calculation unit,

 The incident light conditions are set to 8 kinds, that is, the incident light source is from the upper left, the upper right, the upper right, the right right, the lower right, the lower right, the lower left, and the right left. The difference in brightness between the target pixel and the related pixel typically includes:

 Calculated under the incident light conditions at the upper left:

AL _i j=(L _A ij - LBi-ij-i) X A1 + (L _Ai ,j - L _B ij-i) X A1—1 + (L _Ai ,j - L _Bi-1 ,j) X Al_2 Calculated under incident light conditions directly above:

AL _i j=(L _A ij - L _B ij-i) X A1 + (L _Ai ,j - L _Bi-1 ,j) X Al_l + (L _Ai ,j - L _Bi+1 ,j) X Al_2 Calculated in the upper right incident light condition:

AL _i j=(L _A ij - LBi+ij-i) X A1 + (L _A ij - L _B ij-i) X Al_l + (L _Ai ,j - L _Bi+ ij) Al_2 where A _1j represents an image The destination pixel, that is, the pixel that regulates the brightness, L _Al , j indicates the brightness of the pixel; the central pixel A _1j has a total of 8 pixels around, and the clockwise arrangement from the upper left corner is:

 Bi-i,j-i, Bij_i , Bi+ij_i , Bi+ij, Bi+ij+i , Bij+i , Bi_ij+i , Bi_ij,

 To LB^ indicates the corresponding pixel brightness,

 The incident light conditions are set to 8 kinds, that is, the incident light source is from the upper left, the upper right, the upper right, the right right, the lower right, the lower right, the lower left, and the right left;

(4-3) Pixel color visual perception of brightness performance of the scene space perceptual equivalent D _j calculation unit, under the same direction of observation:

D _LlJ = (L _AlJ / 100)

 Under and under the condition of observing reverse illumination:

D _LlJ = (l- L _Al0 / 100)

 Under conditions of observation and vertical illumination:

Set L _ZH to the target brightness, ie the brightness value for the highest visual perception resolution, value 75-85, if L _Al , >> L _ZH , D _LlJ = (100-L _AlJ ) / (100- L _ZH )

 If LAij < LZH' Duj = LAij / LZH;

(4-4) pixel color visual perception brightness control brightness L _j calculation unit,

 Liij = LAi,j+ ALij X DLij X KL

 Where KL is the set regulation factor, the value is 1.0-2.0;

(4-5) pixel color visual perception hue color gray brightness equivalent D _CAIx and adjusted brightness L _T3⁄4j calculation unit,

Read the image device 0° -359° L _{Cmaxhl of each} color phase plane, and normalize the calculation with the maximum value as the base. The result is the color gray brightness equivalent D _CAIx of the corresponding phase plane.

Compare the pixel color visual perception brightness L _{Ald is} greater than L _Cmaxhl to calculate the adjusted brightness:

 LT2i,j = Lli,j+ ((LAij" LcmaxLl)/( 100- LcmaxLl) ^ DcAIx) ^ Κχ2

Where K _T2 is the set control coefficient, the value range is 0.0-3.0, typically 1.0-2.0, and D _CAIx is the corresponding data of the color phase plane of the pixel color;

(4-6) color gray brightness equivalent D _BAOx and adjusted brightness L _T3⁄4j calculation unit of pixel color visual perception saturation,

Comparing pixel color visual perception brightness L _Al , j is greater than L _Cmaxhl to calculate pixel color visual perception saturation C _Al , j color gray brightness equivalent D _BAOX:

DBAOx ⁼ ( CAij CmaxLl )

c _maxU is the maximum saturation data of the color phase surface brightness sequence of the pixel color, and the adjusted brightness is calculated:

 LT3i,j = LT2i,j+ ((LAij- LcmaxLl)/( 100- LcmaxLl) ^ DBAOX) ^ Κχ3

Where K _T3 is the set control coefficient, the value range is 0.0-3.0, typically 1.0-2.0;

(5) The hue h, the saturation C of the pixel color, and the calculated adjusted brightness L _T3⁄4j or L _j are inversely transformed into the sRGB space R, G, and B values and the normalized calculation module.