基於光纖陣列之快照式超光譜影像系統;Snapshot Hyperspectral Imaging System Base on Fiber Array

Please use this identifier to cite or link to this item: https://ir.lib.ncu.edu.tw/handle/987654321/97662

Title:	基於光纖陣列之快照式超光譜影像系統;Snapshot Hyperspectral Imaging System Base on Fiber Array
Authors:	邱峻嘉;QIU, JUN-JIA
Contributors:	光電科學與工程學系
Keywords:	光纖陣列;快照式超光譜影像;Fiber array;Snap-shot hyperspectral imaging
Date:	2025-07-22
Issue Date:	2025-10-17 11:44:55 (UTC+8)
Publisher:	國立中央大學
Abstract:	隨著時代進步，人們開啟了太空遙測的熱潮，因此發射了許多衛星，而為了使用這些衛星觀察超光譜影像，最初的傳統掃描式超光譜影像系統被開發出來並應用於太空遙測上，此技術對於光譜分析有相當大的潛力，漸漸的超光譜影像被應用於生物分子光譜和螢光光譜觀察上，然而傳統的超光譜影像是利用掃描的方式取得影像資訊再後處理還原回原本的影像，在這個過程中將耗費許多時間，無法有效觀測活體生物，並且在長時間對樣本掃描下，很有可能對樣本本身造成不可逆的光損害。為了實現超光譜系統應用於實時觀察的層面上，以及改良損害樣本的可能性，快照式超光譜影像系統的技術被提出，而本研究旨於搭建出一套對於光譜波長及光譜強度能得到準確數值的快照式超光譜影像系統，搭配陣列光纖，將樣本光切割成細小的子影像，再將所有子影像的光譜蒐集後，便能獲得樣本上每個子影像的光譜曲線，或是將特定波長的所有子影像光譜強度匯集後，產生連續化的超光譜影像，未來經測試成功後將移至實驗室既有的掃描式超光譜螢光系統，將掃描後的螢光利用本系統轉換成實時影像，藉由此系統，我們成功去除掃描的步驟，大大的提升取得超光譜影像的效率，並且成功還原出正確的光譜強度圖，目前由於繞射極限以及非線性色散的影響使長波長波段光譜重疊之情形出現，導致校正完的光譜波長尚未完全準確，未來可以更換元件降低繞射極限的影響來減少光譜重疊，或是更換成線性色散的色散元件使得色散後的光譜波長更容易預測。;As technology advances, hyperspectral imaging has been developed to analyze complex terrains and is now widely applied in molecular spectroscopy. However, traditional hyperspectral imaging relies on scanning to acquire image data, which is then reconstructed through post-processing. This process is time-consuming and makes it difficult to effectively observe living organisms. Moreover, prolonged scanning can potentially cause irreversible photodamage to the samples. To enable real-time observation and reduce the risk of sample damage, snapshot hyperspectral imaging technology has been proposed. This study aims to develop a snapshot hyperspectral imaging system capable of accurately capturing both spectral wavelengths and intensities. By integrating a fiber array, the system splits the light from the sample into small sub-images. The spectral data from each sub-image is collected, allowing for the generation of spectral curves for each region of the sample. Additionally, by compiling the spectral intensity of all sub-images at specific wavelengths, a continuous hyperspectral image can be constructed. Through this system, we eliminate the need for scanning, significantly improving the efficiency of hyperspectral image acquisition, and successfully reconstruct accurate spectral intensity maps. However, due to the influence of the diffraction limit and nonlinear dispersion effects, spectral overlapping still occurs in the long-wavelength region, resulting in minor inaccuracies in wavelength calibration. Future improvements may include replacing optical components to mitigate diffraction limitations or adopting linearly dispersive elements to achieve more predictable spectral separation and enhance calibration accuracy.
Appears in Collections:	[Graduate Institute of Optics and Photonics] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	19	View/Open

社群 sharing

Loading...