Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/2646
標題: 影像處理技術於生物醫學檢測分析上之應用
The Application of Image Processing Techniques in Bio-Medical Detection
作者: 黃富翔
關鍵字: Threshold;二值閥值;Auto-Focus;Edge Detection;Image Expansion;自動對焦;邊緣檢測;影像擴張
出版社: 機械工程學系
摘要: 
本論文針對影像處理之最佳二值閥值、自動對焦技術、邊緣檢測及影像擴張等四個部分分別提出新的方法。
最佳二值化的作法是以影像的像素灰階分佈圖為依據,並以像素統計量的落差及灰階值的落差來作為最佳二值閥值的判斷法則,在本論文的實驗中證明本研究所提出的方法較不易產生誤判的情形。
不同焦距方向所拍攝的模糊影像之最佳二值閥值具有一規則性的變化,即焦距愈模糊時其最佳二值閥值也就會愈低。本論文所提出之自動對焦法便是以此最佳二值閥值作為理論基礎,發展一新的自動對焦技術,其優點為計算式較簡單,佔用的系統資源亦較少。
在邊緣檢測上本論文所提出的二維Lagrange遮罩邊緣檢測法的主要優點乃是此法可以直接作次像素的邊緣檢測,而不必像Canny邊緣檢測法或LoG邊緣檢測法一般,還得再多經一道內插的步驟才能進行次像素邊緣檢測的工作,且所得的結果亦比其他方法來得精確。
最後在影像擴張方面,本論文利用二維三次Lagrange遮罩邊緣檢測法搭配雙邊指數函數進行邊緣修正以減少模糊擴大之情形。
為測試本研究所發展各項影像處理技術之整合性能,我們以步進馬達與精密線性滑軌架設自動對焦平台,再結合CCD組成影像擷取系統,並將所發展之各項新技術以Borland C++ Builder 程式語言撰寫,串成影像檢測系統,最後以組織工程之細胞養殖影像做為測試對象,實驗結果證實本研究所發展之技術確可成功應用於生醫檢測上。

In this thesis, new approaches of the computer vision techniques such as the optimal threshold, the auto focusing, the edge detection, and the image expansion are investigated.
For the optimal threshold, the piecewise searching method that detects the threshold according to the histogram of the gray-level distribution is adopted. Experimental results show that the misjudge problems of the conventional methods is improved. It is also found that the worse the out-of focus is, the lower the optimal threshold is. This physical phenomenon is adopted to develop the new auto focusing method. Less complexity is the main advantage of the proposed auto focusing scheme.
For better edge detection, we use the Lagrange polynomials to build a 2D Lagrange mask such that the required computation time is reduced and the subpixel edge can be directly estimated. The 2D Lagrange edge detection method along with the bipolar sigmoid function is then used to modify the expended edge such that the obscure expansion problem is decreased.
To verify the performance of these new approaches, an auto focusing platform which consists of a linear slider and a step motor, a high resolution couple charged device (CCD), and an image processing software that is written in Borland C++ Builder, are brought together to build a complete image inspection system. The image processing tasks of the cell culture in tissue engineering demonstrate that the proposed image inspection system can be successfully applied to the biomedical analysis problems.
URI: http://hdl.handle.net/11455/2646
Appears in Collections:機械工程學系所

Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.