| 摘要: | 漫反射光譜可以提供物質對於光線的吸收與散射特性的研究資訊，並藉此得知物體組成物質分子的組成成分及顆粒大小，能夠被應用於許多領域之中，如色度量測、配色、太空遙測、植物學及醫學檢測等，然而，傳統的漫反射光譜系統大多僅能得到待測區域的平均光譜資訊，無法得到漫反射光譜在不同空間分布的差異，若能進一步取得光譜的空間資訊，則能得到更多有用的資訊。
不同位置的皮膚因為色素的累積不同，對不同波長的光譜吸收與反射率有所不同，會形成明暗不同的區塊，而以往綜合的光譜無法得到特定空間的反射率資訊，此時若加入光譜的空間資訊，就可以分析皮膚中不同分子，如黑色素、缺氧血紅素、帶氧血紅素等在空間中的分布關係。
因此本研究建構了一套以激發光譜掃描為基礎的漫反射超光譜影像系統，藉由選取不同波長的光源個別打在同一樣本上，記錄物體在不同波長照射下的反射強度分佈影像以獲得物體反射及吸收率的空間分布，建立包含二維空間與一維光譜的超光譜資訊。本研究將建構完成之超光譜系統應用於人體皮膚之漫反射超光譜量測，拍攝人體皮膚漫反射影像，並將量測到的皮膚反射光譜使用理論模型的公式進行分析，得到每種不同樣本區域的含氧量、血紅素的總和濃度以及黑色素的濃度，再使用假設檢定分析實驗所得之數據關係，例如：手背與掌心有血管與沒有血管的區域帶有的含氧量、血紅素總和濃度、黑色素濃度是否具有差異。
若能進一步量測更多數據，形成更龐大的資料庫，則能有助於醫學上對於皮膚疾病的診斷，例如藉由追蹤觀察黑色素濃度的差異，可以判斷黑色素是否異常增生，幫助醫生診斷是否具有黑色素瘤生長的可能，能夠早期發現並治療，提高皮膚癌治癒率。;Diffuse reflectance spectra can provide the information about absorption and scattering characteristics of substance, finding out the tissue composition and the size of granule of objects. This technique can be applied to many fields such as color measurement, color matching, remote sensing, phytology and medical detection. However, in the past, the analysis of diffuse reflectance spectra couldn’t get the difference about the different spatial distribution of diffuse reflectance spectra, but the average spectra information of entire area. If we combine the spatial information with spectra, we can get more useful information.
The biological tissue chromophores at arbitrary locations on skin are variety, which have different absorbance and reflectance depending on wavelength. By combining the spatial information with spectra, we can analyze the correlation between the absorbance and reflectance of the biological tissue chromophores, such as melanin, deoxygenated hemoglobin, oxygenated hemoglobin, β-carotene on skin.
In this article, we set up a multiple-excitation-wavelength system, which are able to select different wavelength exposures to the sample, recording the spatial distribution of absorbance and reflectance. Subsequently, establish hyperspectral images, including 2D spatial information and 1D spectral information. We take diffuse reflectance hyperspectral images of human skin, analyzing them by the formula of theoretical models. Afterwards, get the hemoglobin oxygen saturation, hemoglobin concentration and melanin concentration at different locations of the sample; use hypothesis testing to analyze the experimental data, such as the variability of hemoglobin oxygen saturation, hemoglobin concentration and melanin concentration between the blood and bloodless area on the back and palm of hand.
If we collect more data and accumulate bigger database, it can be useful for the medical diagnosis of skin disease. For instance, investigating the temporal variability of melanin concentration, which can tell whether melanin abnormally grow, can help the doctor to diagnose melanoma. The earlier the melanoma is diagnosed; the higher cure rate of skin cancer is. |
