光學式指紋辨識模組的原理，主要是藉由指紋中未接觸到稜鏡的紋谷，與接觸到稜鏡的紋脊，發生全反射與消散（被吸收）現象，造成光線能量強弱的情形，再經由成像透鏡成像於感測器上，達到擷取指紋影像的功能。 光學設計之初，為了能有效、快速判斷設計架構之正確性，視發生全反射現象的「紋谷」，為數個發光的「點光源」，以此概念建構之模型，簡稱「點近似模型」。事實上會發生全反射現象的紋谷，是「線光源」發光，而非理想點光源發光，因此以線光源發光概念建構之模型，本文簡稱「類指紋模型」。 本文以點近似模型與類指紋模型，在ASAP光機軟體內，分別模擬四種已知指紋辨識模組（傳統式、改良式、新型一、新型二）架構，並探討可辨識的指紋大小、成像畸變量、相對畸變量與影像對比度的差異。 由類指紋模型分別模擬傳統式、改良式、新型一及新型二可得結論如下：在指紋位置+7.5 mm與-7.5 mm時，改良式之影像對比度各為0.65及0.64，傳統式則為0.56及0.32，明顯看出改良式之成像品質優於傳統式。在指紋位置+7.5 mm與-7.5 mm時，新型一之影像對比度各為0.95及0.50，新型二則為0.72及0.90，差異並不明顯，但以整體指紋位置相較，新型二的影像對比度皆在0.72以上，明顯優於新型一。 本文提出以田口方法之直交表，配合國際ISO10110公差標準，進而分析傳統式、改良式、新型一與新型二四種光學模組之對稱非球面鏡與非對稱非球面鏡的允差範圍，是否符合制式規範。藉由田口方法的分析，可知各因子對成像品質影響大小，以做為調整結構參數的依據。最後更進一步修正新型一與新型二的架構，使其在正、負指紋方向的影像對比度更好。 The valleys and ridges of a finger occurs total internal reflection and scattering phenomenon, and then images and focus on the sensor with the imaging lens. According to total internal reflection and scattering result in power of light energy, distinguish out the ridges and valleys of a finger. Finally, the image of fingerprint is captured with a sensor. In order to analyze structure effectively and quickly, we regard the valleys as the point sources. In this thesis, with the model of this concept is called “Approximate Point Model”. Actually, take place valleys of total internal reflection is line sources and is not point sources. Therefore, the concept of line source with the model in this paper is called “Similar Fingerprint Model”. We use Approximate Point Model and Similar Fingerprint Model to simulate four kinds of fingerprint identification by the optical commercial software, ASAP. Four kinds of optical modules are Traditional Type, Modified Type, New Type 1 and New Type 2 respectively. In order to analyze the difference of imaging quality is between Traditional Type, Modified Type, New Type 1 and New Type 2. We use three items to estimate the imaging quality. Three items are distortion, relative distortion and image contrast. The Modified Type is better than Traditional Type. Because of the image contrasts are 0.56 and 0.32 at +7.5 mm and -7.5 mm in Traditional Type, and the image contrasts are 0.65 and 0.64 at +7.5 mm and -7.5 mm in Modified Type. The minimum image contrast is 0.50 in New Type1 and 0.72 in New Type2. According to the result we can know New Type2 is better than New Type1. Design final is tolerance analysis. We first use Taguchi Method and ISO10110 to execute tolerance analysis. We can know relation of each factor influence image contrast. It can help us to correct the value of tolerance. And we modify New Type 1 and New Type 2 to attain better image contrast finally.
