基于室内自然光条件采集舌图像色差校正方法的研究

目的：建立基于室内自然光条件下所采集舌图像的色差校正方法，为舌像颜色准确定量分析奠定方法学基础。方法：将尼康D70 610万像素单镜头反光数码照相机分别设置于自动白平衡（未校正）和手动白平衡（硬件校正）模式，拍摄麦克贝斯24色色卡5种红色类色块和154例舌象，应用NIKON CAPURE NX 1．0．1软件对自动白平衡（未校正）拍摄的图像进行软件校正：通过色差值计算、Imatest软件测试、人工判读比较三类图像（未校正、硬件校正、软件校正）的色差情况。结果：色差值计算结果表明，在自动白平衡（未校正）、手动白平衡（硬件校正）、软件校正三种方法中，软件校正图像的总色差值最小；Imatest软件测试结果提示软件校正图像的色差最小；人工判断结果提示，软件校正的舌像颜色与实际舌象的颜色最为接近，符合率最高。结论：室内自然光条件下采集的舌像通过软件校正后，颜色还原真实。该方法操作方便，有着良好的中医临床应用前景。

Correction of Color Difference in Tongue Image Collection under the Conditions of Indoor Natural Light

Objective： To establish the correction method of color difference in tongue image collection under the conditions of indoor natural light, thus laying the methodological basis for quantitative analysis to tongue image, Methods： The photographs of the red series in the twenty-four GretagMacbeth ColorChecker chart and 154 cases of tongue were taken with the NIKON D70 digital single lens reflex （SLR） camera in the auto white balance mode （uncorrected version）and manual white balance mode（hardware corrected version） respectively. Then the uncorrected version was corrected with the NIKON CAPURE NX 1.0.1 software to get the software corrected ones. The color difference of the three versions of the pictures were compared by the calculation of color difference value, software testing and manual interpretation. Results： Both the results of the calculation of color difference value and the software testing proved that the color difference of the software corrected picture is minimal. The manual interpretation also indicated that with the highest coincidence rate the software corrected picture is close to the actual tongue image mostly. Conclusion： Corrected with software, the tongue pictures taken under the conditions of indoor natural light can represent the tongue image with exactness and fidelity. The method is easy to operate and has a good perspective in clinical application.