自然光条件下基于色差校正方法的舌象颜色分析

目的：尝试建立一种自然光条件下基于L＊a＊b＊色差校正的舌象采集分析方法,观测色差校正方法的舌象颜色分类规律。 方法：在光线良好的室内自然光环境下,应用NikonD70数码单反相机采集413例舌象,以柯达Q-13灰阶色卡为软件校正依据。通过NikonCaptureNX软件进行图像色差校正,并应用L＊a＊b＊均匀色空间对校正后的舌象进行舌象颜色分类与定量分析。 结果：应用软件校正方法对舌象颜色进行色差校正,具有良好的效果。校正舌色后,从淡白舌、淡红舌到红绛舌,L＊值逐渐降低,而a＊值则逐渐升高,差异均有统计学意义（P〈0.01）。红绛舌和青紫舌L＊值的差异无统计学意义,淡红舌和青紫舌a＊值的差异亦无统计学意义。各类舌色的b＊值除淡白舌和青紫舌以及淡红舌和红绛舌之间的差异无统计学意义外,其他舌色两两比较的差异均有统计学意义（P〈0.01）。苔色定量分析显示白苔a＊值大于黄苔（P〈0.01）,白苔b＊值小于黄苔（P〈0.01）。白苔、黄苔L＊值比较,差异无统计学意义。 结论：自然光条件下应用L＊a＊b＊色空间对舌色进行色差校正和定量分析,颜色还原准确,操作方便,是一种可行的临床舌象采集分析方法。

Analysis of tongue color under natural daylight based on chromatic aberration correction

Objective： To establish an analytical method for tongue image acquisition under natural daylight based on L. a ＊ b ＊ error correction, and to observe the classification rules of tongue color using color error correction.
Methods： The tongue images in 413 cases were collected under natural indoor daylight by using Nikon D70 digital SLR camera, and then the color error was adjusted by using Nikon Capture NX software correction according to Kodak Q-13 grey card. The classification and quantitative analysis of the tongue color after software correction was carried out depending on L, a ＊ b ＊ color space.
Results： The software correction method had good effects in adjusting the tongue color image error. The L ＊ values of light red, deep red and cyanosis tongues decreased as compared with that of light white tongue （P〈0.01）, while the a ＊ values of light red, deep red and cyanosis tongues increased as compared with light white tongue （P〈0.01）. There was no significant difference in L＊ value between deep red tongue and cyanosis tongue, and there was also no significant difference in a ＊ value between light red tongue and cyanosis tongue. The b, values of light red and deep red tongues increased as compared with that of light white tongue （P〈0.01）, while the b＊ value of cyanosis tongue decreased as compared with light red and deep red tongues （P〈0. 01）. There was no significant difference in b ＊ value between light white tongue and cyanosis tongue, and there was also no significant difference in b ＊ value between light red tongue and deep red tongue. The a ＊ value of white fur was higher than that of yellow fur （P〈0. 01）, while the b ＊ value of white fur was lower than that of yellow fur （P〈0.01）. There was no significant difference in L＊ value between white fur and yellow fur,
Conclusion： The analytical method for tongue image acquisition under natural daylight based on L ＊ a ＊ b ＊ error correction is accurate in color restoration and feasible to operate.