CN1945627A

CN1945627A - Method for correcting digital tongue picture colour cast

Info

Publication number: CN1945627A
Application number: CN 200610113870
Authority: CN
Inventors: 白净; 张永红; 吴佳; 崔珊珊; 孙晓静
Original assignee: Tsinghua University
Current assignee: Tsinghua University
Priority date: 2006-10-20
Filing date: 2006-10-20
Publication date: 2007-04-11
Anticipated expiration: 2026-10-20
Also published as: CN100412906C

Abstract

The invention relates to a digital correction method for tongue image color shift, which is characterized in that it comprises the following steps in sequence: after performing white balance correction on a Macbeth color calibration color card image taken with a flashlight, it is transferred into hue, saturation, lightness space, and color The standard value of the card is compared to obtain the corresponding adjustment parameters; then use a digital camera to take a picture of the tongue image, then adjust the white balance according to the above method, and then enter the hue, saturation, lightness space, use the adjustment parameters obtained above to adjust the hue, saturation , after the saturation is adjusted, it is converted back to the red, green, and blue space; finally, the gamma value is used to adjust the image in the red, green, and blue space. The gamma value of the curve correction value is fitted by the lightness value of the Macbeth color card standard value and the lightness average value in the color card photo using the least square method as a standard. Tests have proved that this method more accurately reflects the true color of the tongue and has been approved by doctors.

Description

Digital Tongue Image Color Shift Correction Method

技术领域technical field

本发明属于舌象色偏的数字校正方法范畴。The invention belongs to the category of digital correction method for tongue image color shift.

背景技术Background technique

从20世纪80年代中后期开始，中医工作者和数字图像工作者开始联合，尝试以计算机技术为基础进行舌诊客观化研究。研究主要集中在舌头照片颜色的校准，存储和输出和用现代图像处理分析技术对舌头的照片进行处理等方面。其中，舌头的颜色是舌诊中最重要的信息。使用数码相机强制闪光模式拍摄的舌象，虽然能够利用闪光灯的强光线屏蔽自然光线干扰，但是闪光灯的强光照射舌头表面，会使舌象的整体颜色产生偏差，整个舌象偏向红黄的暖色调，因此需要通过一定的方法来校正此偏差。Since the mid-to-late 1980s, Chinese medicine practitioners and digital imaging practitioners began to combine to try to conduct objective research on tongue diagnosis based on computer technology. The research mainly focuses on the color calibration, storage and output of tongue photos, and the processing of tongue photos with modern image processing and analysis techniques. Among them, the color of the tongue is the most important information in tongue diagnosis. Although the strong light of the flash can shield the interference of natural light in the tongue image taken with the forced flash mode of a digital camera, the strong light of the flash illuminates the surface of the tongue, which will make the overall color of the tongue image deviate, and the whole tongue image tends to be warm red and yellow. Hue, so it is necessary to correct this deviation through a certain method.

本方法利用修正数码照片颜色常用的Macbeth24色标准比色卡，该比色卡经过特殊工艺处理，表面具有和皮肤相近的反射光谱，通过照片中Macbeth24比色卡的颜色和真实比色卡颜色的比对，设计出一系列校准参数，用于校准拍摄的舌头图片。This method uses the Macbeth24 color standard color card commonly used to correct the color of digital photos. The color card has been processed by a special process, and the surface has a reflection spectrum similar to the skin. For comparison, a series of calibration parameters are designed to calibrate the tongue pictures taken.

发明内容Contents of the invention

本发明的目的在于校正舌象拍摄时，由强闪光灯造成的色偏，提供一套通用的舌象颜色校正的计算机方法。The purpose of the present invention is to correct the color shift caused by the strong flashlight when the tongue image is taken, and to provide a general computer method for tongue image color correction.

本发明的特征在于：The present invention is characterized in that:

步骤(1)，在强制闪光条件下，用数码相机拍摄N张Macbeth色卡，N＝10～15张，并把照片输入计算机；Step (1), under the forced flash condition, take N pieces of Macbeth color cards with a digital camera, N=10～15 pieces, and photo input computer;

步骤(2)，计算机对N张Macbeth色卡照片校正白平衡：Step (2), the computer corrects the white balance of N Macbeth color card photos:

若：所述N张照片Macbeth色卡照片表示为f₁(x，y)，f₂(x，y)...f_N(x，y)该照片中每一幅图像表示为：f_i(x，y)＝[R_i(x，y)，G_i(x，y)，B_i(x，y)]，i＝1...NR_i(x，y)，G_i(x，y)，B_i(x，y)分别代表在采样点(x，y)的红色、绿色、蓝色的值；则校正白平衡后各照片相应的图片表示为：If: the N photos of the Macbeth color card photos are expressed as f ₁ (x, y), f ₂ (x, y)...f _N (x, y) each image in the photos is expressed as: f _i (x, y)=[R _i (x, y), G _i (x, y), B _i (x, y)], i=1...NR _i (x, y), G _i (x , y), B _i (x, y) respectively represent the red, green, and blue values at the sampling point (x, y); then the corresponding pictures of each photo after correcting the white balance are expressed as:

f_i′＝[R_i′(x，y)，G_i′(x，y)，B_i′(x，y)]，f _i '=[R _i '(x, y), G _i '(x, y), B _i '(x, y)],

${R R}_{i i}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max (({R R}_{i i} ((x x,, y the y))))} * * {R R}_{i i} ((x x,, y the y))$

其中， ${G_{i}}^{'} (x, y) = \frac{255}{Max (G_{i} (x, y))} * G_{i} (x, y)$ in, ${G_{i}}^{'} (x, the y) = \frac{255}{Max (G_{i} (x, the y))} * G_{i} (x, the y)$

${B B}_{i i}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max (({B B}_{i i} ((x x,, y the y))))} * * {B B}_{i i} ((x x,, y the y))$

其中，Max(R_i(x，y))，Max(G_i(x，y))和Max(B_i(x，y))分别表示在图像f_i(x，y)中，红色通道、绿色通道和蓝色通道的最大值；Among them, Max(R _i (x, y)), Max(G _i (x, y)) and Max(B _i (x, y)) respectively represent in the image f _i (x, y), the red channel, the maximum value of the green and blue channels;

步骤(3)，把将在RGB空间表示的f_i′(x，y)和Macbeth色卡颜色转入HSV空间表示，HSV分别表示色调，饱和度和明度：Step (3), transfer f _i '(x, y) and Macbeth color card colors expressed in RGB space into HSV space representation, and HSV represents hue, saturation and lightness respectively:

步骤(3.1)，把f_i′(x，y)＝[R_i′(x，y)，G_i′(x，y)，B_i′(x，y)]转换到HSV空间：Step (3.1), convert f _i ′(x, y)=[R _i ′(x, y), G _i ′(x, y), B _i ′(x, y)] to HSV space:

${H h}_{i i}^{' '} ((x x,, y the y)) = = {cos cos}^{- - 11} [[\frac{(({R R}_{i i}^{' '} ((x x,, y the y)) - - {G G}_{i i}^{' '} ((x x,, y the y)))) + + (({R R}_{i i}^{' '} ((x x,, y the y)) - - {B B}_{i i}^{' '} ((x x,, y the y))))}{22 \sqrt{{(({R R}_{i i}^{' '} ((x x,, y the y)) - - {G G}_{i i}^{' '} ((x x,, y the y))))}^{22} + + (({R R}_{i i}^{' '} ((x x,, y the y)) - - {B B}_{i i}^{' '} ((x x,, y the y)))) * * (({G G}_{i i}^{' '} ((x x,, y the y)) - - {B B}_{i i}^{' '} ((x x,, y the y))))}}]]$

${S S}_{i i}^{' '} ((x x,, y the y)) = = \frac{Max Max (({R R}_{i i}^{' '} ((x x,, y the y)),, {G G}_{i i}^{' '} ((x x,, y the y)),, {B B}_{i i}^{' '} ((x x,, y the y)))) - - Min Min (({R R}_{i i}^{' '} ((x x,, y the y)),, {G G}_{i i}^{' '} ((x x,, y the y)),, {B B}_{i i}^{' '} ((x x,, y the y))))}{Max Max (({R R}_{i i}^{' '} ((x x,, y the y)),, {G G}_{i i}^{' '} ((x x,, y the y)),, {B B}_{i i}^{' '} ((x x,, y the y))))}$

V_i′(x，y)＝Max(R_i′(x，y)，G_i′(x，y)，B_i′(x，y))V _i '(x, y) = Max(R _i '(x, y), G _i '(x, y), B _i '(x, y))

所述H_i′(x，y)，S_i′(x，y)，V_i′(x，y)分别表示转换后第i幅图像在采样点(x，y)的色调、饱和度、明度三个通道的值，而R_i′(x，y)，G_i′(x，y)，B_i′(x，y)分别表示调整白平衡后第i幅图像在采样点(x，y)的红、绿、蓝三个通道的值，校正白平衡后图像的HSV通道表示为：The H _i ′(x, y), S _i ′(x, y), and V _i ′(x, y) represent the hue, saturation, saturation, The values of the three channels of lightness, and R _i ′(x, y), G _i ′(x, y), and B _i ′(x, y) respectively represent the i-th image at the sampling point (x, y) after adjusting the white balance. The values of the red, green and blue channels of y), the HSV channel of the image after correcting the white balance is expressed as:

f_i′(x，y)＝[H_i′(x，y)，S_i′(x，y)，V_i′(x，y)]f _i '(x, y) = [H _i '(x, y), S _i '(x, y), V _i '(x, y)]

步骤(3.2)，把Macbeth色卡的标准值M_j转入HSV空间，Step (3.2), transfer the standard value _Mj of the Macbeth color card into the HSV space,

M_j＝[MR_j，MG_j，MB_j]，j＝1～24M _j ＝[MR _j , MG _j , MB _j ], j=1～24

${MH MH}_{j j} = = {cos cos}^{- - 11} [[\frac{(({MR MR}_{j j} - - {MG MG}_{j j})) + + (({MR MR}_{j j} - - {MB MB}_{j j}))}{22 \sqrt{{(({MR MR}_{j j} - - {MG MG}_{j j}))}^{22} + + (({MR MR}_{j j} - - {MG MG}_{j j})) (({MR MR}_{j j} - - {MB MB}_{j j}))}}]]$

则 ${MS}_{j} = \frac{Max ({MR}_{j}, {MG}_{j}, {MB}_{j}) - Min ({MR}_{j}, {MG}_{j}, {MB}_{j})}{Max ({MR}_{j}, {MG}_{j}, {MB}_{j})}$ but ${MS}_{j} = \frac{Max ({MR}_{j}, {MG}_{j}, {MB}_{j}) - Min ({MR}_{j}, {MG}_{j}, {MB}_{j})}{Max ({MR}_{j}, {MG}_{j}, {MB}_{j})}$

MV_j＝Max(MR_j，MG_j，MB_j)MV _j = Max(MR _j , MG _j , MB _j )

其中，MR_j，MG_j，MB_j分别表示色卡第j个色块的红、绿、蓝三个通道的值，Among them, MR _j , MG _j , and MB _j respectively represent the values of the red, green and blue channels of the jth color block of the color card,

MH_j，MS_j，MV_j分别表示转换后第j个色块的色调、饱和度、明度三个通道的值；MH _j , MS _j , and MV _j represent the values of the three channels of hue, saturation, and lightness of the jth color block after conversion;

得到：色卡标准值的HSV空间表达方式：M_j＝[MH_j，MS_j，MV_j]，j＝1～24；Obtained: the HSV space expression of the standard value of the color card: M _j = [MH _j , MS _j , MV _j ], j = 1~24;

步骤(4)，按照以下步骤通过标准Macbeth色卡与色卡照片的值分别求出各调整参数：In step (4), according to the following steps, each adjustment parameter is obtained respectively through the values of the standard Macbeth color card and the color card photo:

步骤(4.1)，求出调整白平衡后Macbeth色卡照片色调H通道的平均值，并把其与标准色卡平均色调作比较，得到色调调整参数h：Step (4.1), calculate the average value of the color tone H channel of the Macbeth color card photo after adjusting the white balance, and compare it with the average color tone of the standard color card to obtain the tone adjustment parameter h:

N副Macbeth照片的色调平均值为 $Hav = \frac{1}{N} Σ_{i = 1}^{N} {Hav}_{i},$ Hav_i为调整白平衡后的第i个图像的色调H通道的平均值，The average color tone of N Macbeth photos is $Hav = \frac{1}{N} Σ_{i = 1}^{N} {Hav}_{i},$ Hav _i is the average value of the hue H channel of the i-th image after adjusting the white balance,

Macbeth色卡的标准色调平均值为 $Hst = \frac{1}{24} Σ_{j = 1}^{24} {MH}_{j},$ The standard hue average of the Macbeth color card is $Hst = \frac{1}{twenty four} Σ_{j = 1}^{twenty four} {M H}_{j},$

色调调整参数为h＝Hav-Hst；The hue adjustment parameter is h=Hav-Hst;

步骤(4.2)，求出调整白平衡后Macbeth色卡照片明度V通道的平均值，并把其与标准色卡平均明度作比较，得到明度调整参数v：Step (4.2), calculate the average value of the lightness V channel of the Macbeth color card photo after adjusting the white balance, and compare it with the average lightness of the standard color card to obtain the lightness adjustment parameter v:

N副Macbeth照片的明度平均值为 $Vav = \frac{1}{N} Σ_{i = 1}^{N} {Vav}_{i},$ Vav_i为调整白平衡后的第i个图像的明度V通道的平均值，The average brightness of N Macbeth photos is $Vav = \frac{1}{N} Σ_{i = 1}^{N} {Vav}_{i},$ Vav _i is the average value of the lightness V channel of the i-th image after adjusting the white balance,

Macbeth色卡的标准明度平均值为 $Vst = \frac{1}{24} Σ_{j = 1}^{24} {MV}_{j},$ The standard lightness average of the Macbeth color card is $Vst = \frac{1}{twenty four} Σ_{j = 1}^{twenty four} {MV}_{j},$

明度调整参数为v＝Vav-Vst；Brightness adjustment parameter is v=Vav-Vst;

步骤(4.3)，求调整自平衡后Macbeth色卡照片饱和度S通道的平均值，并把其与色卡标准值的饱和度做比较，通过拟合得到饱和度调整参数s：Step (4.3), calculate the average value of the saturation S channel of the photo of the Macbeth color card after adjusting the self-balancing, and compare it with the saturation of the standard value of the color card, and obtain the saturation adjustment parameter s by fitting:

在N张Macbeth色卡照片中，色块j的饱和度平均值为 ${Sav}_{j} = \frac{1}{N} Σ_{i = 1}^{N} {fS}_{ij},$ 其中，fS_ij为第i幅照片中色块j的饱和度，i＝1～N，j＝1～24，In the N Macbeth color card photos, the average saturation of color patch j is ${Sav}_{j} = \frac{1}{N} Σ_{i = 1}^{N} f_{ij},$ Among them, fS _ij is the saturation of color block j in the i-th photo, i=1～N, j=1～24,

色卡中各色块的标准饱和度为MS₁，MS₂...MS₂₄，The standard saturation of each color block in the color card is MS ₁ , MS ₂ ... MS ₂₄ ,

饱和度调整参数s为： $s = \frac{Σ_{j = 1}^{24} {Sav}_{j} * {MS}_{j}}{Σ_{j = 1}^{24} {Sav}_{j}^{2}};$ The saturation adjustment parameter s is: $the s = \frac{Σ_{j = 1}^{twenty four} {Sav}_{j} * {MS}_{j}}{Σ_{j = 1}^{twenty four} {Sav}_{j}^{2}};$

步骤(5)，求出曲线校正值gamma：Step (5), calculate the curve correction value gamma:

在N张Macbeth色卡照片中，色块j的明度值为 ${Vav}_{j} = \frac{1}{N} Σ_{i = 1}^{N} {fV}_{ij},$ 其中，fV_ij为第i幅照片中色块j的明度，i＝1～N，j＝1～24，In the N Macbeth color card photos, the lightness value of the color block j is ${Vav}_{j} = \frac{1}{N} Σ_{i = 1}^{N} f_{ij},$ Among them, fV _ij is the lightness of the color block j in the i-th photo, i=1～N, j=1～24,

色卡中各色块的标准明度为Vav₁，Vav₂...Vav₂₄，The standard lightness of each color block in the color card is Vav ₁ , Vav ₂ ...Vav ₂₄ ,

曲线校正值gamma为： $gamma = \frac{Σ_{j = 1}^{24} \ln ({Vav}_{j}) * \ln ({MV}_{j})}{Σ_{j = 1}^{24} \ln {({Vav}_{j})}^{2}};$ The curve correction value gamma is: $gamma = \frac{Σ_{j = 1}^{twenty four} \ln ({Vav}_{j}) * \ln ({MV}_{j})}{Σ_{j = 1}^{twenty four} \ln {({Vav}_{j})}^{2}};$

步骤(6)，用数码相机拍摄舌头照片，得到图像It(x，y)＝[IR(x，y)，IG(x，y)，IB(x，y)]，其中，IR(x，y)，IG(x，y)，IB(x，y)分别代表在该图像采样点(x，y)红色、绿色、蓝色的值；步骤(7)，校正舌象照片It(x，y)中的白色为纯白色，得到校正后的舌象照片为It′(x，y)：Step (6), take a picture of the tongue with a digital camera to obtain an image It (x, y)=[IR (x, y), IG (x, y), IB (x, y)], wherein, IR (x, y) y), IG(x, y), and IB(x, y) respectively represent the values of red, green, and blue at the image sampling point (x, y); step (7), correct the tongue image photo It(x, The white in y) is pure white, and the corrected tongue picture is It′(x, y):

It′(x，y)＝[IR′(x，y)，IG′(x，y)，IB′(x，y)]It'(x,y)=[IR'(x,y), IG'(x,y), IB'(x,y)]

${IR IR}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max ((IR IR ((x x,, y the y))))} * * IR IR ((x x,, y the y))$

${IG IG}_{i i}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max ((IG IG ((x x,, y the y))))} * * IG IG ((x x,, y the y))$

${IB IB}_{i i}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max ((IB IB ((x x,, y the y))))} * * IB IB ((x x,, y the y))$

其中，Max(IR(x，y))，Max(IG(x，y))和Max(IB(x，y))分别表示在图像It(x，y)中，红色通道、绿色通道和蓝色通道的最大值；Among them, Max(IR(x, y)), Max(IG(x, y)) and Max(IB(x, y)) respectively represent the red channel, green channel and blue channel in the image It(x, y). The maximum value of the color channel;

步骤(8)，将在RGB空间表示的舌象图片转换到用色调H，饱和度S，明度V空间表示Step (8), the tongue image picture represented in RGB space is converted to use hue H, saturation S, lightness V space representation

${IH IH}^{' '} ((x x,, y the y)) = = {cos cos}^{- - 11} [[\frac{(({IR IR}^{' '} ((x x,, y the y)) - - {IG IG}^{' '} ((x x,, y the y)))) + + (({IR IR}^{' '} ((x x,, y the y)) - - {IB IB}^{' '} ((x x,, y the y))))}{22 \sqrt{{(({IR IR}^{' '} ((x x,, y the y)) - - {IG IG}^{' '} ((x x,, y the y))))}^{22} + + (({IR IR}^{' '} ((x x,, y the y)) - - {IB IB}^{' '} ((x x,, y the y)))) (({IG IG}^{' '} ((x x,, y the y)) - - {IB IB}^{' '} ((x x,, y the y))))}}]]$

${IS IS}^{' '} ((x x,, y the y)) = = \frac{Max Max (({IR IR}^{' '} ((x x,, y the y)),, {IG IG}^{' '} ((x x,, y the y)),, {IB IB}^{' '} ((x x,, y the y)))) - - Min Min (({IR IR}^{' '} ((x x,, y the y)),, {IG IG}^{' '} ((x x,, y the y)),, {IB IB}^{' '} ((x x,, y the y))))}{Max Max (({IR IR}^{' '} ((x x,, y the y)),, {IG IG}^{' '} ((x x,, y the y)),, {IB IB}^{' '} ((x x,, y the y))))}$

IV′(x，y)＝Max(IR′(x，y)，IG′(x，y)，IB′(x，y))；IV'(x,y)=Max(IR'(x,y), IG'(x,y), IB'(x,y));

步骤(9)，判断在校正白平衡后，舌象图片曝光是否过度，对曝光不足的图片要进行明度均衡：Step (9), after correcting the white balance, determine whether the tongue picture is overexposed, and perform brightness balance on the underexposed picture:

步骤(9.1)，对所述的照片做明度直方图分布：Step (9.1), the brightness histogram distribution is done to the photo:

直方图HistV(t)＝Num(IV′(x，y)＝t)，t＝0～255，表示在直方图中，第t个分量代表的意思为在明度图IV′(x，y)中值为t的点的个数；Histogram HistV(t)=Num(IV'(x, y)=t), t=0~255, which means that in the histogram, the tth component means that in the brightness map IV'(x, y) The number of points whose median value is t;

步骤(9.2)，求明度图的峰值PeakV：Step (9.2), find the peak value PeakV of the lightness map:

PeakV＝Max(HistV(t))，t＝0～255；PeakV=Max(HistV(t)), t=0～255;

步骤(9.3)，判断是否PeakV＜thresholdVStep (9.3), judge whether PeakV<thresholdV

其中，thresholdV为设定值取值100，若PeakV≥thresholdV，则对IV′(x，y)不处理，Wherein, thresholdV is the setting value of 100, if PeakV≥thresholdV, then IV'(x, y) will not be processed,

否则，进入下一个步骤；Otherwise, go to the next step;

步骤(9.4)，对图像IV′(x，y)作直方图均衡化，得到 ${IV}^{'} (x, y) = Σ_{k = 1}^{{IV}^{'} (x, y)} HistV ({IV}^{'} (x, y)),$ k表示明度值的计数器；Step (9.4), perform histogram equalization on the image IV'(x, y), and obtain ${IV}^{'} (x, the y) = Σ_{k = 1}^{{IV}^{'} (x, the y)} HistV ({IV}^{'} (x, the y)),$ k represents the counter of the lightness value;

步骤(10)，根据步骤(4)设定的调整参数h、s、v对HSV空间内的图像进行校正：Step (10), the image in the HSV space is corrected according to the adjustment parameters h, s, v set in step (4):

IH_d(x，y)＝IH′(x，y)-hIH _d (x, y) = IH'(x, y)-h

IS_d(x，y)＝IS′(x，y)*s；IS _d (x, y) = IS'(x, y)*s;

IV_d(x，y)＝IV′(x，y)-vIV _d (x, y) = IV'(x, y) - v

步骤(11)，将HSV空间内的图像转化到RGB空间内：Step (11), convert the image in the HSV space into the RGB space:

步骤(11.1)，设定四个临时变量f，aa，bb，cc帮助转化：Step (11.1), set four temporary variables f, aa, bb, cc to help transform:

其中，f＝IH_d-floor(IH_d)，函数floor(IH_d)的意思是取比IH_d小的最大的整数，故f表示的是IH_d的小数部分，Wherein, f=IH _d -floor(IH _d ), the meaning of function floor(IH _d ) is to get the largest integer smaller than IH _d , so what f represents is the fractional part of IH _d ,

aa＝IV_d*(1-IS_d)aa=IV _d *(1-IS _d )

bb＝IV_d*(1-(IS_d*f))bb＝IV _d *(1-(IS _d *f))

cc＝IV_d*(1-(IS_d*(1-f)))cc=IV _d *(1-(IS _d *(1-f)))

步骤(11.2)，根据IH_d的范围确定IR_d，IG_d，IB_d的值：Step (11.2), determine the value of IR _d , IG _d , IB _d according to the range of IH _d :

若 ${IH}_{d} &Element; [0 ~ \frac{π}{6}),$ 则IR_d＝IV_d，IG_d＝cc，IB_d＝aalike ${IH}_{d} &Element; [0 ~ \frac{π}{6}),$ Then IR _d =IV _d , IG _d =cc, IB _d =aa

若 ${IH}_{d} &Element; [\frac{π}{6} ~ \frac{π}{3}),$ 则IR_d＝bb，IG_d＝IV_d，IB_d＝aalike ${IH}_{d} &Element; [\frac{π}{6} ~ \frac{π}{3}),$ Then IR _d = bb, IG _d = IV _d , IB _d = aa

若 ${IH}_{d} &Element; [\frac{π}{3} ~ \frac{π}{2}),$ 则IR_d＝aa，IG_d＝IV_d，IB_d＝cclike ${IH}_{d} &Element; [\frac{π}{3} ~ \frac{π}{2}),$ Then IR _d = aa, IG _d = IV _d , IB _d = cc

若 ${IH}_{d} &Element; [\frac{π}{2} ~ \frac{2 π}{3}),$ 则IR_d＝aa，IG_d＝bb，IB_d＝IV_d like ${IH}_{d} &Element; [\frac{π}{2} ~ \frac{2 π}{3}),$ Then IR _d = aa, IG _d = bb, IB _d = IV _d

若 ${IH}_{d} &Element; [\frac{2 π}{3} ~ \frac{5 π}{6}),$ 则IR_d＝cc，IG_d＝aa，IB_d＝IV_d like ${IH}_{d} &Element; [\frac{2 π}{3} ~ \frac{5 π}{6}),$ Then IR _d =cc, IG _d =aa, IB _d =IV _d

若 ${IH}_{d} &Element; [\frac{5 π}{6} ~ 2 π),$ 则IR_d＝IV_d，IG_d＝aa，IB_d＝bblike ${IH}_{d} &Element; [\frac{5 π}{6} ~ 2 π),$ Then IR _d =IV _d , IG _d =aa, IB _d =bb

步骤(12)，用曲线校正值gamma对图像进行调整：Step (12), adjust the image with the curve correction value gamma:

IR_d′(x，y)＝IR_d(x，y)^gamma IR _d '(x, y) = IR _d (x, y) ^gamma

IG_d′(x，y)＝IG_d(x，y)^gamma IG _d '(x, y) = IG _d (x, y) ^gamma

IB_d′(x，y)＝IB_d(x，y)^gamma IB _d '(x, y) = IB _d (x, y) ^gamma

得到的调整后图像It_d(x，y)＝[IR_d′(x，y)，IG_d′(x，y)，IB_d′(x，y)]就是调整后的舌象照片。The obtained adjusted image It _d (x, y)=[IR _d ′(x, y), IG _d ′(x, y), IB _d ′(x, y)] is the adjusted tongue image photo.

使用结果证明经过医生目测，确定调整后的图象比未调整的图象更确切的反映了舌头的真实颜色，得到医生的认可。The result of use proves that the adjusted image reflects the true color of the tongue more accurately than the unadjusted image after the doctor's visual inspection, which is recognized by the doctor.

附图说明Description of drawings

图1：算法流程图；Figure 1: Algorithm flow chart;

图2：Macbeth色卡实物图；Figure 2: The physical picture of the Macbeth color card;

图3：Macbeth色卡示意图；Figure 3: Schematic diagram of the Macbeth color card;

图4：Macbeth色卡真值表；Figure 4: Macbeth color card truth table;

图5：照片的直方图：(a)曝光未足的照片明度直方图，(b)作直方图均衡化之后的明度直方图；Figure 5: The histogram of the photo: (a) the brightness histogram of the underexposed photo, (b) the brightness histogram after histogram equalization;

图6：拍摄的一张舌象照片；Figure 6: A photo of the tongue image taken;

图7：调整白平衡、明度均匀化之后的舌象照片；Figure 7: A photo of the tongue image after white balance adjustment and brightness uniformization;

图8：调整完的舌象照片结果。Figure 8: The result of the adjusted tongue photo.

具体实施方式Detailed ways

该方法硬件由一个数码照相机和一台计算机组成。调整算法的软件由Matlab实现。输入的舌头照片由中医大夫用数码相机拍摄，拍摄时病人平伸出舌头，方便医生将整个舌头摄入照片。The method hardware consists of a digital camera and a computer. The software for adjusting the algorithm is realized by Matlab. The imported tongue photo is taken by a doctor of traditional Chinese medicine with a digital camera. When taking the photo, the patient sticks out his tongue, which is convenient for the doctor to take the whole tongue into the photo.

Macbeth色卡由24个色块组成，分布为四行六列，按照从左至右，从上到下的顺序排列，如图2所示。依次表示为M_j＝[MR_j，MG_j，MB_j]，j＝1～24，顺序示意图如图3所示。其中，MR_j，MG_j，MB_j分别表示第j个色块中红色，绿色，蓝色通道的数值，其附带说明中列有该卡24个色块的标准值，如图4所示。The Macbeth color card consists of 24 color blocks, distributed in four rows and six columns, arranged in order from left to right and from top to bottom, as shown in Figure 2. The sequence is expressed as M _j =[MR _j , MG _j , MB _j ], j=1-24, and the schematic diagram of the sequence is shown in FIG. 3 . Among them, MR _j , MG _j , and MB _j represent the values of the red, green, and blue channels in the jth color block respectively, and the standard values of the 24 color blocks of the card are listed in the accompanying instructions, as shown in Figure 4.

采用尼(Sony)Cybershot DSC-W1数码相机，在室内拍照，拍摄时避免强光干扰，同时使用强制闪光灯模式，距离为30厘米左右。拍摄10张Macbeth色卡，将照片输入到计算机中。Use Sony Cybershot DSC-W1 digital camera to take pictures indoors, avoid strong light interference when shooting, and use forced flash mode at the same time, the distance is about 30 cm. Take 10 Macbeth color cards and import the photos into the computer.

对十幅图像调整白平衡，将图像中最亮的颜色调整到(255，255，255)，其他颜色作相应的线性调整。再将其转化到HSV空间。Adjust the white balance of the ten images, adjust the brightest color in the image to (255, 255, 255), and make corresponding linear adjustments for other colors. Then convert it to HSV space.

对于转化到HSV空间中的图像，与标准色卡颜色做对比，求出参量。在我们的实验中，得到四个参量分别是h＝0.02，s＝-5，v＝-6，gamma＝1.1。For the image converted into HSV space, compare it with the color of the standard color card to find the parameters. In our experiment, the four parameters obtained are h=0.02, s=-5, v=-6, gamma=1.1.

摄入一张舌头的照片，存入计算机。如图6所示。对该照片调整白平衡后，转换到HSV空间。求V通道的直方图，发现该图像属于曝光不足类型，对图像进行明度均衡化，得到图片如图7所示。Take a picture of your tongue and save it to your computer. As shown in Figure 6. After adjusting the white balance of this photo, convert to HSV space. Find the histogram of the V channel, and find that the image is underexposed, and equalize the brightness of the image, as shown in Figure 7.

再按照这套系统得到的参数h＝0.02，s＝-5，v＝-6，gamma＝1.1对图像进行相应的调整，最后调整后的结果如图8所示。According to the parameters h=0.02, s=-5, v=-6, gamma=1.1 obtained by this system, the image is adjusted accordingly, and the final adjusted result is shown in FIG. 8 .

本发明具体实施步骤如下：The specific implementation steps of the present invention are as follows:

1.拍摄N张Macbeth色卡1. Take N pieces of Macbeth color cards

拍摄条件为：在室内拍照，拍摄时避免强光干扰，同时使用强制闪光灯模式，距离为30厘米左右。拍摄N张Macbeth色卡，N＝10～15。将照片输入到计算机中。The shooting conditions are: take pictures indoors, avoid strong light interference when shooting, and use forced flash mode at the same time, with a distance of about 30 cm. Take N pieces of Macbeth color cards, N=10~15. Import photos into your computer.

2.对这N张Macbeth色卡照片校正白平衡2. Correct the white balance of these N Macbeth color card photos

将储存在计算机中的N张照片表示为f₁(x，y)，f₂(x，y)...f_N(x，y)，其中，对于每一幅图像，f_i(x，y)＝[R_i(x，y)，G_i(x，y)，B_i(x，y)]，i＝1...N。(x，y)代表图像的采样点，R_i(x，y)，G_i(x，y)，B_i(x，y)分别代表在该采样点红色、绿色、蓝色的值。Denote the N photos stored in the computer as f ₁ (x, y), f ₂ (x, y)...f _N (x, y), where, for each image, f _i (x, y)=[R _i (x, y), G _i (x, y), B _i (x, y)], i=1...N. (x, y) represents the sampling point of the image, and R _i (x, y), G _i (x, y), and Bi ₍ x, y) represent the red, green, and blue values at the sampling point, respectively.

由于在数码照片中的纯白色的红、绿、蓝三个颜色的值分别为255，255，255，是图像中最亮的颜色，也是各通道的最大值。为了校正拍摄图像的亮度色偏，在校正图像颜色之前，需要将照片中的白色校正为纯白色，其方法为：Since the pure white red, green, and blue colors in digital photos have values of 255, 255, and 255 respectively, they are the brightest color in the image and the maximum value of each channel. In order to correct the brightness and color cast of the captured image, before correcting the image color, it is necessary to correct the white in the photo to pure white, the method is:

${R R}_{i i}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max (({R R}_{i i}))} * * {R R}_{i i} ((x x,, y the y))$

${G G}_{i i}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max (({G G}_{i i}))} * * {G G}_{i i} ((x x,, y the y))$

${B B}_{i i}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max (({B B}_{i i}))} * * {B B}_{i i} ((x x,, y the y))$

其中，Max(R_i)，Max(G_i)和Max(B_i)表示在图像f_i(x，y)中，红色通道、绿色通道和蓝色通道的最大值。调整后相应图片表示为f_i′＝[R_i′(x，y)，G_i′(x，y)，B_i′(x，y)]。通过调整白平衡，我们得到一系列调整后的图像f₁′(x，y)，f₂′(x，y)...f_N′(x，y)。Wherein, Max(R _i ), Max(G _i ) and Max(B _i ) represent the maximum values of the red channel, the green channel and the blue channel in the image f _i (x, y). The adjusted corresponding picture is expressed as f _i '=[R _i '(x, y), G _i '(x, y), B _i '(x, y)]. By adjusting the white balance, we obtain a series of adjusted images f ₁ ′(x,y), f ₂ ′(x,y) . . . f _N ′(x,y).

3.将在RGB(红、绿、蓝)空间表示的f_i′(x，y)和色卡颜色M_j＝[MR_j，MG_j，MB_j]，j＝1～24转入HSV(色调，饱和度，明度)空间表示3. The f _i '(x, y) and the color card color M _j =[MR _j , MG _j , MB _j ] expressed in RGB (red, green, blue) space, j=1～24 are transferred to HSV ( hue, saturation, lightness) spatial representation

对于已经校正白平衡的图像f_i′(x，y)＝[R_i′(x，y)，G_i′(x，y)，B_i′(x，y)]，通过下式将其转换到HSV空间中。For the image f _i '(x, y)=[R _i '(x, y), G _i '(x, y), B _i '(x, y)] for the image f i '(x, y)] whose white balance has been corrected, it is expressed by the following formula Convert to HSV space.

其中，R_i′(x，y)，G_i′(x，y)，B_i′(x，y)分别表示调整白平衡后第i幅图像在采样点(x，y)的红、绿、蓝三个通道的值，H_i′(x，y)，S_i′(x，y)，V_i′(x，y)分别表示转换后第i幅图像在采样点(x，y)的色调、饱和度、明度三个通道的值。由此得到校正白平衡之后图像的HSV通道表示为Among them, R _i ′(x, y), G _i ′(x, y), and B _i ′(x, y) respectively represent the red and green colors of the i-th image at the sampling point (x, y) after adjusting the white balance. , the values of the three channels of blue, H _i ′(x, y), S _i ′(x, y), V _i ′(x, y) respectively represent the i-th image after conversion at the sampling point (x, y) The values of the three channels of Hue, Saturation, and Brightness. Thus, the HSV channel of the image after correcting the white balance is expressed as

另外，将对色卡的标准值M_j＝[MR_j，MG_j，MB_j]，j＝1～24也转入到HSV空间。In addition, the standard value M _j =[MR _j , MG _j , MB _j ] of the color card, j=1~24 is also transferred to the HSV space.

${MS MS}_{j j} = = \frac{Max Max (({MR MR}_{j j},, {MG MG}_{j j},, {MB MB}_{j j})) - - Min Min (({MR MR}_{j j},, {MG MG}_{j j},, {MB MB}_{j j}))}{Max Max (({MR MR}_{j j},, {MG MG}_{j j},, {MB MB}_{j j},,))}$

MV_j＝Max(MR_j，MG_j，MB_j)MV _j = Max(MR _j , MG _j , MB _j )

其中，MR_j，MG_j，MB_j分别表示色卡第j个色块的红、绿、蓝三个通道的值，MH_j，MS_j，MV_j分别表示转换后第j个色块的色调、饱和度、明度三个通道的值。从而得到色卡标准值的HSV空间表达方式M_j＝[MH_j，MS_j，MV_j]，j＝1～24。Among them, MR _j , MG _j , and MB _j represent the values of the red, green, and blue channels of the jth color block of the color card respectively, and MH _j , MS _j , and MV _j represent the hue of the jth color block after conversion , Saturation, Brightness three channel values. Thus, the HSV space expression of the standard value of the color card is obtained M _j =[MH _j , MS _j , MV _j ], j=1-24.

4.求出调整白平衡后Macbeth色卡照片H(色调)通道的平均值，与色卡平均色调做比较，得到色调调整参数h4. Find the average value of the H (hue) channel of the Macbeth color card photo after adjusting the white balance, and compare it with the average hue of the color card to obtain the hue adjustment parameter h

用Hav₁，Hav₂，...Hav_N表示调整后的各个图像i的H通道的平均值，则N副图像的色调平均值为 $Hav = \frac{1}{N} Σ_{i = 1}^{N} {Hav}_{i} \cdot$ 另外，求色卡标准色调值的平均值 $Hst = \frac{1}{24} Σ_{j = 1}^{24} {MH}_{j} \cdot$ 两者的差h＝Hav-Hst为调整色度的参数。Use Hav ₁ , Hav ₂ , ... Hav _N to represent the average value of the H channel of each image i after adjustment, then the average value of the tone of the N sub-images is $Hav = \frac{1}{N} Σ_{i = 1}^{N} {Hav}_{i} &Center Dot;$ In addition, find the average value of the standard hue value of the color card $Hst = \frac{1}{twenty four} Σ_{j = 1}^{twenty four} {M H}_{j} \cdot$ The difference between the two h=Hav-Hst is a parameter for adjusting chromaticity.

5.求出调整白平衡后Macbeth色卡照片V(明度)通道的平均值，与色卡平均明度做比较，得到明度调整参数v5. Find the average value of the V (lightness) channel of the Macbeth color card photo after adjusting the white balance, and compare it with the average lightness of the color card to obtain the lightness adjustment parameter v

用Vav₁，Vav₂，...Vav_N表示调整后的各个图像的V通道的平均值，则N副图像的明度平均值为 $Vav = \frac{1}{N} Σ_{i = 1}^{N} {Vav}_{i} \cdot$ 另外，求色卡标准值的平均值 $Vst = \frac{1}{24} Σ_{j = 1}^{24} {MV}_{j} \cdot$ 两者的差v＝Vav-Vst为调整明度的参数。Use Vav ₁ , Vav ₂ ,...Vav _N to represent the average value of the V channels of each image after adjustment, then the average value of the brightness of the N images is $Vav = \frac{1}{N} Σ_{i = 1}^{N} {Vav}_{i} \cdot$ In addition, find the average value of the standard value of the color card $Vst = \frac{1}{twenty four} Σ_{j = 1}^{twenty four} {MV}_{j} &Center Dot;$ The difference v=Vav-Vst between the two is a parameter for adjusting brightness.

6.求出调整白平衡后Macbeth色卡照片S(饱和度)通道的平均值，与色卡的饱和度做比较，通过拟合得到饱和度调整参数s6. Find the average value of the S (saturation) channel of the Macbeth color card photo after adjusting the white balance, compare it with the saturation of the color card, and obtain the saturation adjustment parameter s by fitting

在一张调整了白平衡之后的照片f_i′(x，y)中，按照位置分别取出24个矩形的色块，如图3所示，求得各个色块内饱和度分别为fS_i1，fS_i2....fS_i24。对采集的所有照片都分别求出24个色块的饱和度之后，再求各个色块对于这N张图片的平均值。In a photo f _i ′(x, y) after adjusting the white balance, 24 rectangular color blocks are taken out according to their positions, as shown in Figure 3, and the saturation in each color block is obtained as fS _i1 , fS _i2 .... fS _i24 . After calculating the saturation of 24 color blocks for all the collected photos, calculate the average value of each color block for the N pictures.

${Sav Sav}_{j j} = = \frac{11}{N N} {Σ Σ}_{i i = = 11}^{N N} {fS f}_{ij ij} \cdot &Center Dot;$

对于求得的饱和度平均值Sav₁，Sav₂...Sav₂₄和色卡标准饱和度MS₁，MS₂...MS₂₄，用最小二乘法标准进行拟和，具体方法为：选取参数s，MS_j＝s*Sav_j，使得 $δ = \sqrt{Σ_{j = 1}^{24} {(s * {Sav}_{j} - {MS}_{j})}^{2}}$ 的值最小。For the calculated saturation averages Sav ₁ , Sav ₂ ...Sav ₂₄ and color card standard saturation MS ₁ , MS ₂ ...MS ₂₄ , use the least square method to fit the standard, the specific method is: select the parameters s, MS _j = s*Sav _j , such that $δ = \sqrt{Σ_{j = 1}^{twenty four} {(the s * {Sav}_{j} - {MS}_{j})}^{2}}$ The value of is the smallest.

s的计算方法为： $s = \frac{Σ_{j = 1}^{24} {Sav}_{j} * {MS}_{j}}{Σ_{j = 1}^{24} {Sav}_{j}^{2}}$ The calculation method of s is: $the s = \frac{Σ_{j = 1}^{twenty four} {Sav}_{j} * {MS}_{j}}{Σ_{j = 1}^{twenty four} {Sav}_{j}^{}}$

7.求出调整白平衡后色卡照片V(明度)通道的平均值，与色卡的明度做比较，通过拟合得到曲线校正值gamma7. Find the average value of the V (lightness) channel of the color card photo after adjusting the white balance, compare it with the lightness of the color card, and obtain the curve correction value gamma by fitting

gamma源于CRT(显示器/电视机)的响应曲线，即其明度与输入电压的非线性关系；Gamma comes from the response curve of CRT (monitor/TV), that is, the nonlinear relationship between its brightness and input voltage;

gamma校正相当于对图像进行处理；gamma的变化带来明度的变化，故通过明度V通道来得到gamma的值：Gamma correction is equivalent to processing the image; the change of gamma brings the change of brightness, so the value of gamma is obtained through the brightness V channel:

在一张调整了白平衡之后的照片f_i′(x，y)中，按照位置分别取出24个矩形的色块，求得各个色块内明度数值分别为fV_i1，fV_i2....fV_i24。对采集的所有照片都分别求出24个色块的明度之后，再求各个色块对于这N张图片的平均值。In a photo f _i ′(x, y) after adjusting the white balance, 24 rectangular color blocks are taken out according to the positions, and the lightness values in each color block are obtained as fV _i1 , fV _i2 ... fV _i24 . After calculating the lightness of 24 color blocks for all the collected photos, calculate the average value of each color block for the N pictures.

${Vav Vav}_{j j} = = \frac{11}{N N} {Σ Σ}_{i i = = 11}^{N N} {fV f}_{ij ij} \cdot \cdot$

对于求得的明度平均值Vav₁，Vav₂...Vav₂₄和色卡标准明度MV₁，MV₂...MV₂₄，用最小二乘法标准进行拟和，具体方法为：选取参数gamma，MV_j＝Vav_j ^gamma，使得 $σ = \sqrt{Σ_{j = 1}^{24} {({Vav}_{j}^{gamma} - {MV}_{j})}^{2}}$ 的值最小。For the obtained lightness averages Vav ₁ , Vav ₂ ... Vav ₂₄ and color card standard lightness MV ₁ , MV ₂ ... MV ₂₄ , use the least square method to fit the standard, the specific method is: select the parameter gamma, MV _j = Vav _j ^gamma , such that $σ = \sqrt{Σ_{j = 1}^{twenty four} {({Vav}_{j}^{gamma} - {MV}_{j})}^{2}}$ The value of is the smallest.

gamma的计算方法为 $gamma = \frac{Σ_{j = 1}^{24} \ln ({Vav}_{j}) * \ln ({MV}_{j})}{Σ_{j = 1}^{24} \ln {({Vav}_{j})}^{2}} \cdot$ The calculation method of gamma is $gamma = \frac{Σ_{j = 1}^{twenty four} \ln ({Vav}_{j}) * \ln ({MV}_{j})}{Σ_{j = 1}^{twenty four} \ln {({Vav}_{j})}^{2}} &Center Dot;$

8.用数码相机拍摄舌头照片，存入计算机。8. Take photos of the tongue with a digital camera and store them in the computer.

其表示方法为It(x，y)＝[IR(x，y)，IG(x，y)，IB(x，y)]，(x，y)代表图像的采样点，IR(x，y)，IG(x，y)，IB(x，y)分别代表在该采样点红色、绿色、蓝色的值。Its representation method is It (x, y) = [IR (x, y), IG (x, y), IB (x, y)], (x, y) represents the sampling point of the image, IR (x, y ), IG(x, y), and IB(x, y) respectively represent the values of red, green, and blue at the sampling point.

9.校正舌象照片白平衡9. Correct the white balance of the tongue image photo

将舌象照片中的白色校正为纯白色，其方法为：Correct the white in the tongue image photo to pure white, the method is:

${IR IR}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max ((IR IR))} * * IR IR ((x x,, y the y))$

${IG IG}_{i i}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max ((IG IG))} * * IG IG ((x x,, y the y))$

${IB IB}_{i i}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max ((IB IB))} * * IB IB ((x x,, y the y))$

其中，Max(IR)，Max(IG)和Max(IB)表示在图像It(x，y)中，红色通道、绿色通道和蓝色通道的最大值。通过调整白平衡，我们得到调整后的舌象图片为It′(x，y)。Among them, Max(IR), Max(IG) and Max(IB) represent the maximum values of the red channel, green channel and blue channel in the image It(x, y). By adjusting the white balance, we get the adjusted tongue image as It′(x, y).

10.将在RGB空间表示It_i′(x，y)转入HSV(色调，饱和度，明度)空间表示10. Convert the representation It _i '(x, y) in RGB space to HSV (hue, saturation, lightness) space representation

IV′(x，y)＝Max(IR′(x，y)，IG′(x，y)，IB′(x，y))IV'(x,y)=Max(IR'(x,y),IG'(x,y),IB'(x,y))

11.曝光不足的照片进行明度均衡11. Lightness equalization for underexposed photos

校正白平衡后的舌象，首先要分析明度直方图分布，如果峰值出现的位置小于事先确定的阈值，对舌象作标记，经过明度均衡，再进行下一步的调节。To correct the tongue image after white balance, first analyze the lightness histogram distribution, if the position of the peak value is smaller than the predetermined threshold, mark the tongue image, after lightness equalization, then proceed to the next step of adjustment.

统计图像的明度通道IV′(x，y)的直方图。明度通道的取值为0～255。统计明度通道的直方图HistV(t)，t＝0～255。A histogram of the Luminance channel IV'(x,y) of the statistical image. The value of the lightness channel is 0~255. The histogram HistV(t) of statistical brightness channel, t=0～255.

直方图的定义为HistV(t)＝Num(IV′(x，y)＝t)，即直方图的第t个分量代表的意思为在明度图IV′(x，y)中，值为t的点的个数。求明度图的峰值PeakV＝Max(HistV(t))，t＝0～255，图5(a)表示为一个图像的直方图。横坐标为灰度值，取值从0到255，纵坐标为图像中灰度值等于横坐标灰度的像素点的个数。The histogram is defined as HistV(t)=Num(IV'(x, y)=t), that is, the tth component of the histogram means that in the lightness map IV'(x, y), the value is t The number of points of . Calculate the peak value PeakV=Max(HistV(t)) of the lightness map, t=0～255, and Fig. 5(a) is shown as a histogram of an image. The abscissa is the gray value, ranging from 0 to 255, and the ordinate is the number of pixels in the image whose gray value is equal to the gray value of the abscissa.

如果PeakV＜thresholdV，则对图像IV′(x，y)作直方图均衡化。均衡化的公式表示为： ${IV}^{'} (x, y) = Σ_{k = 1}^{{IV}^{'} (x, y)} HistV ({IV}^{'} (x, y)),$ 表示的意思是，一个像素点(x，y)在均衡化之前的灰度是IV′(x，y)，那么这一点在均衡化之后的明度值是直方图从灰度值1到灰度值IV′(x，y)的和。式子中的k表示明度值，图5(b)为改图通过直方图均衡变化后的直方图。If PeakV<thresholdV, perform histogram equalization on image IV'(x, y). The equalization formula is expressed as: ${IV}^{'} (x, the y) = Σ_{k = 1}^{{IV}^{'} (x, the y)} HistV ({IV}^{'} (x, the y)),$ It means that the gray level of a pixel (x, y) before equalization is IV'(x, y), then the brightness value of this point after equalization is the histogram from gray value 1 to gray Sum of values IV'(x,y). The k in the formula represents the lightness value, and Figure 5(b) is the histogram after the modified image has been changed through histogram equalization.

如果PeakV≥thresholdV，对IV′(x，y)不进行任何处理。If PeakV≥thresholdV, no processing is performed on IV'(x, y).

12.根据设定的参数h、s、v对HSV空间内的图像进行校正；12. Correct the image in the HSV space according to the set parameters h, s, v;

IH_d(x，y)＝IH′(x，y)-hIH _d (x, y) = IH'(x, y)-h

IS_d(x，y)＝IS′(x，y)*sIS _d (x, y) = IS'(x, y)*s

IV_d(x，y)＝IV′(x，y)-vIV _d (x, y) = IV'(x, y) - v

13.将在HSV空间中的图像转化到RGB空间后再用gamma系数13. Convert the image in HSV space to RGB space and then use the gamma coefficient

将HSV转到RGB空间的程序为：The program to convert HSV to RGB space is:

设四个临时变量f，aa，bb，cc来帮助运算，其中Set four temporary variables f, aa, bb, cc to help the operation, where

f＝IH_d-floor(IH_d)，函数floor(IH_d)的意思是取比IH_d小的最大的整数，故f表示的是IH_d的小数部分f=IH _d -floor(IH _d ), the function floor(IH _d ) means to take the largest integer smaller than IH _d , so f represents the fractional part of IH _d

aa＝IV_d*(1-IS_d)aa=IV _d *(1-IS _d )

bb＝IV_d*(1-(IS_d*f))bb＝IV _d *(1-(IS _d *f))

cc＝IV_d*(1-(IS_d*(1-f)))cc=IV _d *(1-(IS _d *(1-f)))

根据IH_d的范围可以确定出IR_d，IG_d，IB_d的值The value of IR _d , IG _d , and IB _d can be determined according to the range of IH _d

在转入到RGB空间后，再用曲线校正值gamma值对RGB空间的图象进行调整。After transferring to the RGB space, adjust the image in the RGB space with the gamma value of the curve correction value.

IR_d′(x，y)＝IR_d(x，y)^gamma IR _d '(x, y) = IR _d (x, y) ^gamma

IG_d′(x，y)＝IG_d(x，y)^gamma IG _d '(x, y) = IG _d (x, y) ^gamma

IB_d′(x，y)＝IB_d(x，y)^gamma IB _d '(x, y) = IB _d (x, y) ^gamma

Claims

1. The digital tongue image color shift correction method is characterized in that it contains the following steps in sequence:

Step (1), under the forced flash condition, take N pieces of Macbeth color cards with a digital camera, N=10～15 pieces, and photo input computer;

Step (2), the computer corrects the white balance of N Macbeth color card photos:

If: the N photos of Macbeth color cards are expressed as f ₁ (x, y), f ₂ (x, y)...f _N (x, y)

Each image in the photo is expressed as: f _i (x, y)=[R _i (x, y), G _i (x, y), B _i (x, y)], i=1...N

R _i (x, y), G _i (x, y), Bi ₍ x, y) respectively represent the red, green, and blue values at the sampling point (x, y);

Then the corresponding picture of each photo after correcting the white balance is expressed as:

f _i '=[R _i '(x, y), G _i '(x, y), B _i '(x, y)],

{R R}_{i i}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max (({R R}_{i i} ((x x,, y the y))))} * * {R R}_{i i} ((x x,, y the y))

in,

{G_{i}}^{'} (x, the y) = \frac{255}{Max (G_{i} (x, the y))} * G_{i} (x, the y)

{B B}_{i i}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max (({B B}_{i i} ((x x,, y the y))))} * * {B B}_{i i} ((x x,, y the y))

Among them, Max(R _i (x, y)), Max(G _i (x, y)) and Max(B _i (x, y)) respectively represent in the image f _i (x, y), the red channel, the maximum value of the green and blue channels;

Step (3), transfer f _i '(x, y) and Macbeth color card colors expressed in RGB space into HSV space representation, and HSV represents hue, saturation and lightness respectively:

Step (3.1), convert f _i ′(x, y)=[R _i ′(x, y), G _i ′(x, y), B _i ′(x, y)] to HSV space:

{H h}_{i i}^{' '} ((x x,, y the y)) = = {cos cos}^{- - 11} [[\frac{(({R R}_{i i}^{' '} ((x x,, y the y)) - - {G G}_{i i}^{' '} ((x x,, y the y)))) + + (({R R}_{i i}^{' '} ((x x,, y the y)) - - {B B}_{i i}^{' '} ((x x,, y the y))))}{22 \sqrt{{(({R R}_{i i}^{' '} ((x x,, y the y)) - - {G G}_{i i}^{' '} ((x x,, y the y))))}^{22} + + (({R R}_{i i}^{' '} ((x x,, y the y)) - - {B B}_{i i}^{' '} ((x x,, y the y)))) * * (({G G}_{i i}^{' '} ((x x,, y the y)) - - {B B}_{i i}^{' '} ((x x,, y the y))))}}]]

{S S}_{i i}^{' '} ((x x,, y the y)) = = \frac{Max Max (({R R}_{i i}^{' '} ((x x,, y the y)),, {G G}_{i i}^{' '} ((x x,, y the y)),, {B B}_{i i}^{' '} ((x x,, y the y)))) - - Min Min (({R R}_{i i}^{' '} ((x x,, y the y)),, {G G}_{i i}^{' '} ((x x,, y the y)),, {B B}_{i i}^{' '} ((x x,, y the y))))}{Max Max (({R R}_{i i}^{' '} ((x x,, y the y)),, {G G}_{i i}^{' '} ((x x,, y the y)),, {B B}_{i i}^{' '} ((x x,, y the y))))}

V _i '(x, y) = Max(R _i '(x, y), G _i '(x, y), B _i '(x, y))

The H _i ′(x, y), S _i ′(x, y), and V _i ′(x, y) represent the hue, saturation, saturation, The values of the three channels of lightness, and R _i ′(x, y), G _i ′(x, y), and B _i ′(x, y) respectively represent the i-th image at the sampling point (x, y) after adjusting the white balance. The values of the red, green and blue channels of y), the HSV channel of the image after correcting the white balance is expressed as:

f _i '(x, y) = [H _i '(x, y), S _i '(x, y), V _i '(x, y)]

Step (3.2), transfer the standard value _Mj of the Macbeth color card into the HSV space,

M _j ＝[MR _j , MG _j , MB _j ], j=1～24

{MH MH}_{j j} = = {cos cos}^{- - 11} [[\frac{(({MR MR}_{j j} - - {MG MG}_{j j})) + + (({MR MR}_{j j} - - {MB MB}_{j j}))}{22 \sqrt{{(({MR MR}_{j j} - - {MG MG}_{j j}))}^{22} + + (({MR MR}_{j j} - - {MG MG}_{j j})) (({MR MR}_{j j} - - {MB MB}_{j j}))}}]]

but

{MS}_{j} = \frac{Max ({MR}_{j}, {MG}_{j}, {MB}_{j}) - Min ({MR}_{j}, {MG}_{j}, {MB}_{j})}{Max ({MR}_{j}, {MG}_{j}, {MB}_{j})}

MV _j = Max(MR _j , MG _j , MB _j )

Among them, MR _j , MG _j , and MB _j respectively represent the values of the red, green and blue channels of the jth color block of the color card,

MH _j , MS _j , and MV _j represent the values of the three channels of hue, saturation, and lightness of the jth color block after conversion;

Obtained: the HSV space expression of the standard value of the color card: M _j = [MH _j , MS _j , MV _j ], j = 1~24;

In step (4), according to the following steps, each adjustment parameter is obtained respectively through the values of the standard Macbeth color card and the color card photo:

Step (4.1), calculate the average value of the color tone H channel of the Macbeth color card photo after adjusting the white balance, and compare it with the average color tone of the standard color card to obtain the tone adjustment parameter h:

The average color tone of N Macbeth photos is

Hav = \frac{1}{N} Σ_{i = 1}^{N} {Hav}_{i},

Hav _i is the average value of the hue H channel of the i-th image after adjusting the white balance,

The standard hue average of the Macbeth color card is

Hst = \frac{1}{twenty four} Σ_{j = 1}^{twenty four} {M H}_{j},

The hue adjustment parameter is h=Hav-Hst;

Step (4.2), calculate the average value of the lightness V channel of the Macbeth color card photo after adjusting the white balance, and compare it with the average lightness of the standard color card to obtain the lightness adjustment parameter v:

The average brightness of N Macbeth photos is

Vav = \frac{1}{N} Σ_{i = 1}^{N} {Vav}_{i},

Vav _i is the average value of the lightness V channel of the i-th image after adjusting the white balance,

The standard lightness average of the Macbeth color card is

Vst = \frac{1}{twenty four} Σ_{j = 1}^{twenty four} {MV}_{j},

Brightness adjustment parameter is v=Vav-Vst;

Step (4.3), find the average value of the saturation S channel of the Macbeth color card photo after adjusting the white balance, and compare it with the saturation of the standard value of the color card, and obtain the saturation adjustment parameter s by fitting:

In the N Macbeth color card photos, the average saturation of color patch j is

{Sav}_{j} = \frac{1}{N} Σ_{i = 1}^{N} f S_{ij,}

Among them, fS _ij is the saturation of color block j in the i-th photo, i=1～N, j=1～24,

The standard saturation of each color block in the color card is MS ₁ , MS ₂ ... MS ₂₄ ,

The saturation adjustment parameter s is:

the s = \frac{Σ_{j = 1}^{twenty four} {Sav}_{j} * {MS}_{j}}{Σ_{j = 1}^{twenty four} {Sav}_{j}^{2}};

Step (5), calculate the curve correction value gamma:

In the N Macbeth color card photos, the lightness value of the color block j is

{Vav}_{j} = \frac{1}{N} Σ_{i = 1}^{N} f V_{ij},

Among them, fV _ij is the lightness of the color block j in the i-th photo, i=1～N, j=1～24,

The standard lightness of each color block in the color card is Vav ₁ , Vav ₂ ...Vav ₂₄ ,

The curve correction value gamma is:

\frac{Σ_{j = 1}^{twenty four} \ln ({Vav}_{j}) * \ln ({MV}_{j})}{Σ_{j = 1}^{twenty four} \ln {({Vav}_{j})}^{2}};

Step (6), take a picture of the tongue with a digital camera to obtain an image It (x, y)=[IR (x, y), IG (x, y), IB (x, y)], wherein, IR (x, y) y), IG(x, y), and IB(x, y) respectively represent the values of red, green, and blue at the image sampling point (x, y); step (7), correct the tongue image photo It(x, The white in y) is pure white, and the corrected tongue picture is It′(x, y):

It'(x,y)=[IR'(x,y), IG'(x,y), IB'(x,y)]

{IR IR}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max ((IR IR ((x x,, y the y))))} * * IR IR ((x x,, y the y))

{I I {G G}_{i i}}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max ((IG IG ((x x,, y the y))))} * * IG IG ((x x,, y the y))

{I I {B B}_{i i}}^{' '} ((x x,, y the y)) = = \frac{255255}{Max Max ((IB IB ((x x,, y the y))))} * * IB IB ((x x,, y the y))

Among them, Max(IR(x, y)), Max(IG(x, y)) and Max(IB(x, y)) respectively represent the red channel, green channel and blue channel in the image It(x, y). The maximum value of the color channel;

Step (8), the tongue image picture represented in RGB space is converted to use hue H, saturation S, lightness V space representation

I I {H h}^{' '} ((x x,, y the y)) = = {cos cos}^{- - 11} [[\frac{(({IR IR}^{' '} ((x x,, y the y)) - - {IG IG}^{' '} ((x x,, y the y)))) + + (({IR IR}^{' '} ((x x,, y the y)) - - {IB IB}^{' '} ((x x,, y the y))))}{22 \sqrt{{(({IR IR}^{' '} ((x x,, y the y)) - - {IG IG}^{' '} ((x x,, y the y))))}^{22} + + (({IR IR}^{' '} ((x x,, y the y)) - - {IB IB}^{' '} ((x x,, y the y)))) * * ((I I {G G}^{' '} ((x x,, y the y)) - - {IB IB}^{' '} ((x x,, y the y))))}}]]

{I I S S}^{' '} ((x x,, y the y)) = = \frac{Max Max (({I I R R}^{' '} ((x x,, y the y)),, {IG IG}^{' '} ((x x,, y the y)),, {IB IB}^{' '} ((x x,, y the y)))) - - Min Min (({IR IR}^{' '} ((x x,, y the y)),, {I I G G}^{' '} ((x x,, y the y)),, {I I B B}^{' '} ((x x,, y the y))))}{Max Max (({IR IR}^{' '} ((x x,, y the y)),, {I I G G}^{' '} ((x x,, y the y)),, {I I B B}^{' '} ((x x,, y the y))))}

IV'(x,y)=Max(IR'(x,y), IG'(x,y), IB'(x,y));

Step (9), after correcting the white balance, determine whether the tongue picture is overexposed, and perform brightness balance on the underexposed picture:

Step (9.1), the brightness histogram distribution is done to the photo:

Histogram HistV(t)=Num(IV'(x, y)=t), t=0~255, which means that in the histogram, the tth component means that in the brightness map IV'(x, y) The number of points whose median value is t;

Step (9.2), find the peak value PeakV of the lightness map:

PeakV=Max(HistV(t)), t=0～255;

Step (9.3), judge whether PeakV<thresholdV

Wherein, thresholdV is the setting value of 100, if PeakV≥thresholdV, then do not process IV'(x, y), otherwise, enter the next step;

Step (9.4), perform histogram equalization on the image IV'(x, y), and obtain

{IV}^{'} (x, the y) = Σ_{k = 1}^{{IV}^{'} (x, the y)} HistV ({IV}^{'} (x, the y)),

k represents the counter of the lightness value;

Step (10), the image in the HSV space is corrected according to the adjustment parameters h, s, v set in step (4):

IH _d (x, y) = IH'(x, y)-h

IS _d (x, y) = IS'(x, y)*s;

IV _d (x, y) = IV'(x, y) - v

Step (11), convert the image in the HSV space into the RGB space:

Step (11.1), set four temporary variables f, aa, bb, cc to help transform:

Wherein, f=IH _d -floor(IH _d ), the meaning of function floor(IH _d ) is to get the largest integer smaller than IH _d , so what f represents is the fractional part of IH _d ,

aa=IV _d *(1-IS _d )

bb＝IV _d *(1-(IS _d *f))

cc=IV _d *(1-(IS _d *(1-f)))

Step (11.2), determine the value of IR _d , IG _d , IB _d according to the range of IH _d :

like

{IH}_{d} &Element; [0 ~ \frac{π}{6}),

Then IR _d =IV _d , IG _d =cc, IB _d =aa

like

{IH}_{d} &Element; [\frac{π}{6} ~ \frac{π}{3}),

Then IR _d =bb, IG _d =IVd, IB _d =aa

like

{IH}_{d} &Element; [\frac{π}{3} ~ \frac{π}{2}),

Then IR _d =aa, IG _d =IVd, IB _d =cc

like

{IH}_{d} &Element; [\frac{π}{2} ~ \frac{2 π}{3}),

Then IR _d = aa, IG _d = bb, IB _d = IV _d

like

{IH}_{d} &Element; [\frac{2 π}{3} ~ \frac{5 π}{6}),

Then IR _d =cc, IG _d =aa, IB _d =IV _d

like

{IH}_{d} &Element; [\frac{5 π}{6} ~ 2 π),

Then IR _d =IV _d , IG _d =aa, IB _d =bb

Step (12), adjust the image with the curve correction value gamma:

IR _d '(x, y) = IR _d (x, y) ^gamma

IG _d '(x, y) = IG _d (x, y) ^gamma

IB _d '(x, y) = IB _d (x, y) ^gamma

The obtained adjusted image It _d (x, y)=[IR _d ′(x, y), IG _d ′(x, y), IB _d ′(x, y)] is the adjusted tongue image photo.