隨著數位影像技術的迅速發展，數位相機在各個商業及工業應用領域中被廣泛使用，這些相機每天拍攝並儲存大量的數位影像。然而，影像在拍攝過程中時常不可避免地受到通訊干擾或影像處理過程中所引起的雜訊影響，並被納入最終資料中。本論文提出一種基於去噪與去模糊方法的創新模型，並整合盲Richardson-Lucy去卷積演算法，以有效解決雜訊影像問題，提升影像品質。
本研究採用U-Net架構並藉由執行上述去模糊模組以促進影像的自適應去噪功能。經過逐一架構及進行實驗驗證後，確認本論文所提出之方法確能有效去除影像之雜訊，並能顯著提升輸出影像之品質。本論文之方法在多種場景下，均呈現良好輸出影像之品質，為模糊影像處理技術提供研究發展之基礎。特別針對環境及數位化雜訊之顯著降低成效，本論文之方法取得影像清晰度及一致性確實達到明確改善成果，將AI理論應用於影像處理技術之研究產生顯著成效。

As digital image technology advances rapidly, digital cameras used in various domains in commercial and industrial applications are taking and storing large amounts of digital images daily. While taking the images, it is inevitable to enclose noises generated by communication disruptions or image processing into the final data. This thesis proposes a novel model based on denoising and deblurring methods embedded with blind Richardson-Lucy deconvolution algorithm to circumvent the noisy images for better quality.
In the thesis, the U-Net architecture is implemented to facilitate adaptive image denoising functions by executing the above-mentioned deblurring modules. Experimental results verify that the proposed approach significantly eliminates noise, leading to a noticeable improvement in image quality. The proposed method demonstrates strong efficacy across diverse scenarios, thereby laying robust groundwork for further development of image processing technologies. The noises could be notably diminished with the method being implemented, resulting in improvement in image clarity and consistency with far-reaching impacts on the future of image processing technologies in realistic applications.