本研究主要在架構一套人眼光學量測系統，測量人眼之點擴散函數(Point Spread Function, PSF)，並用以完善本研究團隊所建構之人眼眼球模型，使其更加符合真實人眼的表現。 此系統基於資訊影像的完整性(例如眼內散射、漫射)，使用二次通過成像(Double-Pass imaging configuration, DP)的架構，及改良式非對稱性二次通過(Asymmetric Double-Pass, ADP)系統進行量測，以計算出人眼光學系統之調制轉換函數(Modulation Transfer Function, MTF)及相位傳遞函數(Phase Transfer Function, PTF)。再透過降低誤差迭代演算法(Error-Reduction Iterative Algorithm, ER)重建出人眼單次通過之點擴散函數。 本研究分為實驗量測及影像重建兩個階段。先以假體校準系統建立標準量測流程，再以人體實驗得到人眼MTF及PTF。接著以降低誤差迭代演算法代入人眼MTF及PTF重建真實人眼之點擴散函數。 實驗所得之人眼點擴散函數在成像分布及MTF的特性表現上，與現有相關文獻資料的結果大致相符，驗證此量測系統的可行性。未來可改變實驗光源的波長及入射角度，量測不同情境下之人眼點擴散函數，用以修正並增加眼球模型的適用性。 This study focused on the construction of a human-eye optical measurement system to measure the Point Spread Function (PSF) of human eyes. By using the data we can optimize the optical parameters of the human-eye model previously developed in the research group to match the performance of real human eyes. In order to retain the information integrity in the measured images, especially the effects of intraocular scattering and diffusion, Double-Pass (DP) and Asymmetric Double-Pass (ADP) configurations are adopted to measure the double-pass PSFs and calculate the Modulation Transfer Function (MTF) and Phase Transfer Function (PTF) of human eyes. Subsequently, the single-pass PSF of human eyes is reconstructed by an Error-Reduction (ER) iterative algorithm. There are two phases in this study: one is the experimental measurement and the other is the image reconstruction. The artificial-eye experiment served as the system calibration and protocol familiarization. The human-eye experiment provided us the MTF and PTF of human eyes. Then we reconstructed the PSF of real human eyes by introducing the MTF and PTF into the ER iterative algorithm. The image distribution and MTF of the human-eye PSF in our experimental results showed the same trend as the data in the literature, which verified the feasibility of this measurement system. By changing the wavelength and incident angle of the light source, the PSFs of human eyes under different conditions can be measured in the future to modify the human-eye model and enhance the model applicability.
