體外震波碎石應用於醫療領域上已有二十八年歷史，震波碎石機由震波產生器、定位系統、三軸治療床等構成，其中震波產生器、聚焦系統和耦合機制用以擊碎結石，定位系統用以透視結石位置與深度，治療床用以調整對位、瞄準結石，影像中看得見結石、準確對位才能將結石擊碎。
目前手動結石定位程序煩瑣、費時，技術人員培訓養成時間長。結石3D立體定位無法如一般三軸平台採用外加電阻尺、電位計或位置編碼器等方式加以定位，原因是結石在患者體內位置、深度、形狀、大小、均不相同，當治療床Z軸高度改變時，影像出現非線性變化關係，導致定位修正量無法確定，此時最佳定位方式為影像追蹤，結石定位大多採用X光透視，如何自動準確迅速定位，減少患者照射X光時間、輻射劑量為本論文研究目標。本論文採用兩種影像追蹤方式，正化灰階關聯辨識法和幾何圖形比對辨識法完成結石影像自動追蹤定位系統。其中定位功能完全由軟體程式操控安裝容易，毋需加裝任何感測元件，影像追蹤定位功能速度快、準確率高。

Extracorporeal shockwave lithotripsy has been used for medical application over 28 years. An extracorporeal shockwave lithotripter is composed of a shock wave generator, a localization system, and a three-axis treatment table. Shock wave generator, focus system and coupling mechanism are used for stone fragmentation; localization system is used to find stone position and depth; treatment table is used for aligning the stone to the focus point . Stone must be seen and localized precisely to be fragmented.
Manual stone localization needs many procedures and a long time; it also takes a long period of time to train a technician. Stone localization belongs to a kind of 3D localization method, but it can’t be achieved by putting position sensor to determine stone position and depth just like traditional 3D table, because each patient’s stone position、depth、shape and size are different. Especially when treatment table height is changing, there is a nonlinear change relation with the image size on the monitor, which causes uncertain factors to modify table movement distance. To make up for this problem, the best solution is image tracking. Most hospitals use X-ray for stone localization. How to automatically localize stone precisely and reduce X-ray exposure time are the research purposes of this thesis. This thesis investigates two image tracking methods, Normalize gray scale correlation pattern match and Geometric pattern match, to accomplish automatic localization for stone image tracking. Such a localization system is controlled solely by software programs; it is easy to set up or to upgrade old system. There is no need to put any sensor in the original system. Image tracking localization can offer high speed and precise stone localization function.

[1] E. Chan, “Review of Block Matching Based Motion Estimation Algorithms for Video Compression”, IEEE Trans. Electrical and Computer Engineering, Vol.1, pp.151-154, 1993.
[2] M. Delius, “Medical Applications and Bio Effects of Extracorporeal Shock Waves”, pp. 55-72, 1994.
[3] H. Gharavi, and M. Mills , “Block Matching Motion Estimation Algorithms-New Results”, IEEE Trans. Circuits and Systems, Vol.37, pp. 649-651, 1990.
[4] W. Hepp, M. Grünewald, and W. Brendel, “Die Extrakorporale Stoßwellen lithotripsie”, Spektrum der Wissenschaft, pp. 44-53, 1991.
[5] H. K. Tang, T. H. Wu , and Y. T. Lin , “Real-time Object Image Tracking Based on Block-Matching Algorithm”, 1998.
[6] M. Müller, “Experimental Investigations on Focusing of Weak Spherical Shock Waves in Water by Shallow Ellipsoidal Reflectors”, pp. 64-85 , 1987.
[7] O. Wess, and T. Locher, “What Makes a Good Lithotripter? ”, Storz Medical AG, Switzerland, 2000.
[8] Matrox Imaging, Inc., “Matrox Image Library User Guide 7.5”, 2002.
[9] http://www.alacron.com/software/SWPATTERN.htm#
[10] http://www.cee.hw.ac.uk/hipr/html/sobel.html
[11] http://www.gehealthcare.com/usen/xr/edu/index.html
[12] http://www.ismst.com/start.htm
[13] http://www.matrox.com/imaging/products/mil/home.cfm
[14] http://www.unc.edu/%7Eslzhu/research/research.html
[15]http://www.urologiena.com/subspecialisaties/niersteenverbrijzeling-eng.htm#Non-Stone
[16]柳忠文，應用區塊比對法於雙軸平台之穩定行性控制，逢甲大學自動控制工程研究所碩士學位論文，民國93年。
[17]陳昭介，門禁監控系統之影像追蹤，國立中央大學電機工程研究所碩士學位論文，民國93年。
[18]黃芸妮，螢光蛋白質自動辨識及追蹤系統，中原大學醫學工程學系碩士學位論文，民國93年。