紅外光譜影像系統能提供物質分子組成的波長與二維空間資訊，因此在許多領域中被廣泛應用。然而，現行系統大多依賴陣列偵測器或空間掃描式干涉架構來取得光譜影像。這些方法面臨兩大挑戰：紅外光波段的陣列偵測器價格高昂，以及干涉技術架構易受環境擾動影響，這使得紅外光譜影像系統在生活應用與研究推廣上受到限制。 為了解決這些問題，本研究提出了一種單像素成像結合散射介質的多光譜影像系統。該系統利用散射介質將物體的空間與光譜資訊轉化為光斑圖像，並透過單像素成像技術進行記錄與重建。由於光斑圖像有著空間相關性與光譜去相關性的特性，因此系統可藉由反摺積技術成功重建物體影像。為了驗證系統可行性，系統目前先以可見光進行實驗，本研究目前成功以陣列相機拍攝到光斑並還原物體的光譜影像，並成功分離出不同波段的物體的光譜影像，在單像素的架構上已成功還原單色光物體的光譜影像，然而在多色光物體影像的還原中，影像輪廓清晰度仍不及單色光物體的影像。此外目前系統無法使還原影像的大小與原始物體達到1:1比例未來若能提升散射介質後的光強度，或調整系統架構與元件參數，增加影像放大率，將有助於提升結合散射介質的單像素多光譜影像系統(Single-pixel Multispectral Imaging System Based on Scattering Medium，以下簡稱SpMI-SS)的訊雜比與系統還原真實物體大小的能力。 ;Infrared spectral imaging systems provide wavelength and two-dimensional spatial information about the molecular composition of substances, making them widely used in various fields. However, current systems primarily rely on array detectors or spatial scanning interferometric architectures to obtain spectral images. These methods face two major challenges: the high cost of array detectors in the infrared wavelength range and the susceptibility of interferometric architectures to environmental disturbances. These limitations hinder the application of infrared spectral imaging systems in everyday life and their widespread adoption in research. To address these challenges, this study proposes a multispectral imaging system combining single-pixel imaging with scattering media. The system utilizes scattering media to transform the spatial and spectral information of an object into speckle patterns, which are recorded and reconstructed using single-pixel imaging techniques. Due to the spatial correlation and spectral decorrelation characteristics of the speckle patterns, the system can successfully reconstruct object images through deconvolution techniques. To validate the feasibility of the system, initial experiments were conducted using visible light. The study successfully captured speckle patterns with an array camera and reconstructed the spectral images of the objects. Additionally, the spectral images of objects at different wavelengths were effectively separated. The single-pixel imaging setup successfully reconstructed spectral images of monochromatic objects. However, when reconstructing multicolor objects, the image contours were less clear compared to those of monochromatic objects. Future improvements in enhancing the light intensity after passing through scattering media or adjusting the system architecture and component parameters to increase the magnification factor will contribute to improving the signal-to-noise ratio (SNR) of the Single-pixel Multispectral Imaging System Based on Scattering Medium (SpMI-SS) and its ability to restore the true size of the original object.
