本論文主要介紹兩影像間的仿設轉換及其應用。
首先，我們會介紹如何找出一影像的特徵點，並利用梯度下降法來找出另一影像的共軛對，根據找出的共軛對的對數，可以分成以下三種情況：第一種情況為找不到三對共軛對，則無法得到仿射轉換；第二種情況為找到剛好三對共軛對，則可利用克拉瑪法則解出方程並得到仿射轉換；第三種情況為找到三對以上共軛對，則可以利用最小平方法找出最適合的仿射轉換。
最後，我們會討論如果有一個凸起物在一個平坦的背景上，則理論上凸起物的表面與平坦的背景的仿射轉換是不一樣的。所以，我們可以藉由仿射轉換的不同判斷出哪些地方可能是凸起物的表面，而知道凸起所在的地方。

The purpose of this study was to investigate how to find the conjugate pairs and affine transformation between two images.We used gradient descent method to find the conjugate pairs of two images.If there are more than three conjugate pairs of two images ,we can use least squares method by these conjugate pairs to find the affine transformation between two images.Another aim was to find out how to determine where the protrusion was.Since the affine transformation of background is different from the affine transformation of the surface of the protrusion.we used this property to determine where the protrusion was.Then,we did the experiment to verify that the affine transformation of background is different from the affine transformation of the surface of the protrusion,and used the property to determine where the protrusion was.Finally, we concluded that the closer camera is to the background ,the more different the affine transformation of background and surface of the protrusion are ,if the protrusion has fixed height.

[1]范光照(2018)，「自動化光學檢測技術」，科學月刊579期
[2]Douwe Osinga,(2018).Deep Learning Cookbook:Practical
Recipes to Get Started Quickly.O’Reilly.
[3]Ramesh Jain, Rangachar Kasturi, Brian G. Schunck,
(1995).Machine Vision.McGraw-Hill, Inc.
[4]H.L.Royden,(1988).Real Analysis(3rd ed.).MacMillan
Publishing Company.
[5]Richard L. Burden, J. Douglas Faires,(2010).Numerical
Analysis(9th ed.).Cengage Learning.
[6]賴岱佑(2008)，數位影像分析之智慧型監視系統，文魁資訊
[7]S. Friedberg, A. Insel, L. Spence,(2014).Linear
Algebra(4th ed.).Pearson.
[8]Jan R. Magnus,Heinz Neudecker,(2019).Matrix
Differential Calculus with Applications in Statistics
and Econometrics(3rd ed.).WILEY.
[9]Rafael C. Gonzalez,Richard E.Woods,(2002).Digital
Image Processing(2nd ed.).Prentice Hall.
[10]Yalin Bastanlar, Yasemin Yardimci,(2008)."Corner
Validation based on Extracted Corner
Properties".Computer Vision and Image Understanding
112,pp.243-261.