Invisible Patterns, Indiscernible Pigments? Non-Destructive Detection of Painted Cultural Relics Using Hyperspectral Imaging

高光譜成像技術在彩繪文物的檢測中具有廣泛的應用，它是一種非破壞性的分析方法，可以提供文物表面詳細的光譜信息，從而幫助研究人員和保護人員識別顏料、評估文物的保存狀況并支持修復工作。

Hyperspectral imaging technology has extensive applications in the examination of painted cultural relics. As a non-destructive analytical method, it provides detailed spectral information about the surface of artifacts, enabling researchers and conservators to identify pigments, assess preservation conditions, and support restoration efforts.

「顏料識別 / Pigment Identification」

高光譜成像可以用于識別彩繪文物中使用的各種顏料。不同的顏料具有的光譜吸收和反射特性，通過建立顏料光譜數據庫，可以將文物表面的光譜數據與數據庫中的標準光譜進行匹配，從而確定顏料的種類。這種方法尤其適用于那些肉眼難以區分的相似顏料的鑒定。

例如，一個課題組利用高光譜成像技術分析佛教洞穴繪畫的顏料成分，為文物的保護和修復提供科學依據。該研究系統建立了包含24種常見礦物顏料（涵蓋紅、白、黃、藍、綠、黑等色系）的粉末狀態高光譜反射光譜數據庫。

為了更貼近實際繪畫條件，數據庫特別納入了添加動物膠粘合劑（1:1比例）后的顏料光譜，因為研究發現膠粘合劑會顯著改變光譜特征（如：可能導致原始吸收峰消失或產生新的吸收峰）。

Hyperspectral imaging can identify various pigments used in painted cultural relics. Different pigments exhibit unique spectral absorption and reflection characteristics. By establishing a pigment spectral database, the spectral data from artifact surfaces can be matched with standard spectra in the database to determine pigment types. This method is particularly useful for distinguishing visually similar pigments that are difficult to differentiate with the naked eye.

For example, one research team utilized hyperspectral imaging to analyze the pigment composition of Buddhist cave paintings, providing a scientific basis for artifact conservation and restoration. The study systematically established a hyperspectral reflectance database for 24 common mineral pigments (covering red, white, yellow, blue, green, black, and other color systems) in powdered form.

To better approximate actual painting conditions, the database specifically included spectra of pigments mixed with animal glue binder (1:1 ratio), as research showed that binders significantly alter spectral features (e.g., causing original absorption peaks to disappear or generating new ones).

顏料粉末與膠混合后，分別涂布在紙板和石材上

The powder and glue are mixed and applied to the cardboard and the stone.

粉末與粉末混合膠的對比：(a) *色；(b) 沙漠棕褐色；(c) 藍花色；(d) 綠松石綠色

Spectral comparison of powder and powder mixed glue: (a) cinnabar; (b) desert tan color; (c) blue flower color; (d) turquoise green color.

「材料評估 / Material Evaluation」

高光譜成像可用于評估彩繪文物的材料構成。通過分析文物表面的光譜數據，可以識別出不同的材料，如顏料、粘合劑和保護涂層。這對于了解文物的制作工藝和材料來源具有重要意義。

一項針對斯洛文尼亞彩繪蜂巢木板畫的研究，系統性地應用了不用波段的高光譜成像系統。這充分發揮了不同波段的優勢：

VNIR提供高分辨率顏色分布，SWIR對有機黏合劑（如干性油）的化學基團（如酯羰基、亞甲基）振動特征敏感，其典型吸收峰在~1940nm和2300~2330nm區域，成功區分了視覺顏色相似但化學成分不同的區域。

Hyperspectral imaging can assess the material composition of painted cultural relics. By analyzing surface spectral data, different materials such as pigments, binders, and protective coatings can be identified. This is crucial for understanding the craftsmanship and material origins of artifacts.

A study on Slovenian painted beehive panels systematically employed hyperspectral imaging systems across different spectral bands, leveraging the advantages of each:

VNIR provided high-resolution color distribution. SWIR was sensitive to vibrational characteristics of chemical groups (e.g., ester carbonyl, methylene) in organic binders like drying oils, with typical absorption peaks at ~1940 nm and 2300–2330 nm, successfully distinguishing areas with similar visual colors but different chemical compositions.

主成分分析(PCA)結果，左：可見近紅外(VNIR，400-1000nm)波段，中：近紅外(NIR，900-1700nm)波段，右：短波紅外(SWIR，900-2500nm)波段

Principal Component Analysis (PCA) results: left: Visible and Near-Infrared (VNIR, 400-1000 nm) band; center:  Near-Infrared (NIR, 900-1700 nm) band; right: Short-Wave Infrared (SWIR, 900-2500 nm) band.

「繪制圖案分析 / Pattern Analysis」

高光譜成像技術可以用于分析彩繪文物的繪制圖案。通過對高光譜圖像進行處理，如主成分圖像選擇、移動最小二乘法（MLS）和圖像融合等方法，可以提取彩繪文物上的繪制圖案，發現肉眼難以觀察到的隱藏信息，這種方法對于研究古代繪畫技巧和文化具有重要意義。

例如，針對秦陵二號銅車馬表面嚴重受損的"夔(Kuí)龍紋"（存在顏料脫落、銅銹覆蓋和變色三類損傷），研究采用近紅外高光譜成像（900-1700nm）技術。通過分析發現，1450nm附近是顏料邊緣的顯著吸收谷，在此波段下拍攝的高光譜圖像能清晰顯示肉眼不可見的紋飾邊緣，顯著克服了銅銹覆蓋和顏料脫落的干擾。

Hyperspectral imaging can analyze painted patterns on cultural relics. Through processing techniques such as principal component image selection, Moving Least Squares (MLS), and image fusion, hidden details invisible to the naked eye can be extracted. This method is highly valuable for studying ancient painting techniques and cultural significance.

For instance, in examining the severely damaged "Kuilong pattern" on the No. 2 Bronze Chariot of the Qin Mausoleum (affected by pigment flaking, bronze rust coverage, and discoloration), near-infrared hyperspectral imaging (900–1700 nm) was employed. Analysis revealed a significant absorption trough near 1450 nm at pigment edges. Hyperspectral images captured at this wavelength clearly displayed pattern edges obscured by rust or pigment loss, overcoming interference from these factors.

光譜采集區域 / The picture of the gathering area

"夔龍紋"邊緣區域光譜曲線域

The spectral curve of the edge areas in the “Kuilong" pattern

(a)、(b)： 經高通濾波器增強的高光譜圖像，及其邊緣提取結果

(c) 、(d)：經圖像校正后的普通光學圖像，及其邊緣提取結果

(e)：圖(b)與(d)邊緣信息的融合結果

(a), (b): Hyperspectral image enhanced by high-pass filter and its edge extraction result

(c), (d): Corrected conventional optical image and its edge extraction result

(e): Fusion result of edge information from (b) and (d)

「科技賦能，續寫文明 / Technology Empowering Cultural Continuity」

高光譜成像技術，如同一把無損的“文物解碼器"，為彩繪文物的研究與保護開辟了新的視角，在揭示顏料信息、鑒別材料組成、再現歷史紋飾方面展現出強大的能力。

精準探測，源于好的設備。要充分發揮高光譜技術的潛力，穩定可靠、性能優異的設備至關重要。我司提供寬廣波段的高光譜成像儀，可集成至便攜式、實驗室臺架等多種系統，為文化遺產現場與實驗室檢測提供無損、精準、高效的整體解決方案。

隨著數據處理方法、光譜數據庫的完善以及多技術協同應用的深入，這項技術將會在文化遺產保護領域扮演越來越關鍵的角色，為守護和傳承人類瑰寶貢獻重要力量。

Hyperspectral imaging, like a non-destructive "artifact decoder," opens new perspectives for the research and conservation of painted cultural relics. It demonstrates remarkable capabilities in revealing pigment information, identifying material compositions, and reconstructing historical patterns.

Precision detection stems from exceptional equipment. To fully realize the potential of hyperspectral technology, stable, reliable, and high-performance instruments are essential. Our company offers broad-band hyperspectral imagers that can be integrated into portable, laboratory bench, and other systems, providing non-destructive, accurate, and efficient solutions for both on-site and laboratory cultural heritage examinations.

As data processing methods improve, spectral databases expand, and multi-technique collaborations deepen, this technology will play an increasingly critical role in cultural heritage conservation, contributing significantly to the preservation and transmission of human treasures.

案例來源 / Source：

1. Shi, X., Lin, X., Lei, Y., Wu, J., Lv, X., & Zhou, Y. (2024). A study on pigment composition of Buddhist cave paintings based on hyperspectral technology. Materials, 17(21), 5147.

2. Sandak, J., Sandak, A., Legan, L., Retko, K., Kav?i?, M., Kosel, J., Poohphajai, F., Diaz, R. H., Ponnuchamy, V., Sajin?i?, N., Gordobil, O., Tavzes, C., & Ropret, P. (2021). Nondestructive evaluation of heritage object coatings with four hyperspectral imaging systems. Coatings, 11(2), 244.

3. Han, D., Ma, L., Ma, S., & Zhang, J. (2019). Discovery and extraction of surface painted patterns on the cultural relics based on hyperspectral imaging. Journal of Physics: Conference Series, 1237(3), 032028.