本論文介紹光學式非接觸手寫辨識系統，利用CMOS攝影機加上手指書寫軌跡，傳送到處理系統中做處理，以達到手寫辨識的效果。非接觸式手寫辨識系統與傳統觸控手寫辨識系統的不同在於觸控手寫辨識系統需要觸控到顯示螢幕才能判斷出手寫的字元。
CMOS攝影機擷取影像後將對影像進行處理，如中值濾波、侵蝕及膨脹等，並以背景相減法進行影像切割，並判別出指尖位置，最後藉由演算法判斷出手寫的字元。經實驗驗證，本光學式非接觸手寫辨識系統已具有辨識字元的效果，具可行性。

This paper describes the optical non-contact of handwriting recognition system. Use the CMOS camera and fingertip writing trajectory, transmission to the processing system to achieve the handwriting recognition’s result. The traditional handwriting recognition system needs to touch the screen to achieve the characters’ recognition is the most different between the optical non-contact of handwriting recognition system.
We processed for the images after CMOS camera captured, such as median filter, erosion and expansion. And use the principle of background subtraction to do the image segmentation and find the fingertip position, the last by algorithm to recognize the hand-written characters. After experimental verification, the system has effective recognize for characters. It can be feasible in the future.

[1] 粘存賢，“設計與實作一個基於雙攝影機視覺系統的雙點觸控螢幕”，國立中央大學資訊工程學系碩士論文，民國九十八年。
[2] 蔡宗典，“超薄ITO 透明導電膜應用在觸控面板之研究”，國立中央大學光電科學研究所碩士論文，民國九十七年。
[3] “電磁式觸控面板”，博拓國際智權集團。
[Online]. Available:
http://www.portal-ipr.com/index.php/zw/electromagnetic
[4] 羅富仁，“非接觸式電場感應系統”，國立中央大學光電科學研究所碩士論文，民國九十八年。
[5] Cheng Wen and Chin-Hung Huang. “A paperless fax machine with a single-touch panel.” in IEEE Transactions on Consumer Electronics, Vol.54, NO.4, November 2008.
[6] Jianwei Wang and Wei Zhao. “A Method of Color Space Selection for Color Forest Inspection Image Denoising Applications.” in ICACTE, Vol.4, No V4-207, August 2010.
[7] Rafael C. Gonzalez and Richard E. Woods. “Digital Image Processing, 3rd Edition”, 2008.
[8] A. Ford and A. Roberts. “Color space conversions”, 1998.
[9] Li Shuhua and Guo Gaizhi. “ The application of improved HSV color space model in image processing.” in ICFCC, vol.2, NO, V2-10, May 2010.
[10] 吳軒孝，“雷射線結構光應用於物體表面及3D輪廓之精細量測”，國立中央大學光電科學研究所碩士論文，民國九十六年。
[11] 林彥如，“生產線上之影像量測系統”，國立中央大學光電科學研究所碩士論文，民國九十九年。