高光谱成像技术对人体面部和手掌的成像及光谱分析

目的：应用高光谱成像技术对正常人体不同部位图像进行采集,分析这些部位的光谱特征,得到正常值数据,并为该技术用于疾病诊断和中医面诊和手诊奠定基础。方法：使用高光谱成像仪,采集10例（男5女5）健康人的面部和双手掌图像,应用高光谱处理分析软件,对面和手掌划分15个分析区域,统计各分区从450-900 nm每间隔10 nm一个光谱段的各个分析区光强度值和光谱特征,并分析个体特点。结果：在面和手掌的高光谱图像上可以清晰地显示面部器官和手掌的不同部位,以580-830 nm段图像更为清晰,530 nm以下图像杂波较多。面部颧、颧下、鼻尖、眉间、额等部位和四指指丘、大小鱼际部位的反光较强,而眼、眉、嘴角和五指末端等则较弱。面部光谱双侧基本对称,而左右手的对称性在不同的个体上不尽相同。在光谱曲线上可以见到某一或者某些波长处光谱出现突变的细节。结论：高光谱成像技术可以清晰地显示人体面和手的图像,显示各部位（器官）的光谱特征;本文得到的面、手高光谱正常数据和特征将为该技术用于疾病诊断和中医辨证提供参考依据。

Imaging and Spectral Analysis of Human Face and Hand by Hyperspectral Imaging Technology

Objective： Collect the different parts of normal human body image by hyperspectral imaging technology and analyzing spectral characteristics of these parts,to get the normal value data,and lay the foundations for disease diagnosis and Chinese medical examination of face and hand by the technology. Methods： Using hyperspectral imaging instrument,10 cases（ male 5 female 5） healthy human of face and double palm image were collected. The face and hand were divided into 15 analysed regions. The light intensity and spectral characteristics in partition from the 450-900 nm every 10 nm were counted and analysed by the spectral analysis processing software. Results： The facial and hand hyperspectral image can clearly show the different parts and organs. 580-830 nm segment image is clearer,under 530 nm is more noisy. Spectral intensities on cheek,below cheek,nasal tip,glabella,forehead and four fingers hill,hand thenar,hypothenar are higher values than those on eyes,eyebrows,lips and fingers at different wavelengths. The bilateral spectrum of face is mainly symmetric,and the symmetry of left and right hand are not the same on different individuals. The spectrum mutation details at one or some wavelength are seen in spectral curve too. Conclusion： Hyperspectral imaging technology can clearly display the image of the human face and hand and show the spectral characteristics of parts（ organ）. The normal spectrum data and features obtained in this paper will provide reference for the technology used for disease diagnosis and TCM syndrome differentiation.