而在硬體方面,由指紋擷取器擷取出 像素大小的指紋影像,透過串列的方式傳輸至DSP裡面的SDRAM,經過前處理、特徵處理以及比對處理,最後以綠色和紅色的LED燈號來顯示比對成功或是失敗。經過測試,硬體辨識結果與模擬結果大致相同,驗證了所實現指紋辨識系統之可行性。 In this thesis, we implement a fingerprint identification system using a digital signal processor (DSP). First, we perform computer simulations on the adopted fingerprint verification algorithm. Two sources of fingerprint images were tested, the first set is adopted from the database of Fingerprint Verification Competition 2000(FVC2000). The second set is acquired by the sensor LTT-C500 manufactured by LighTuning Tech. Inc. Because of its fillister design which puts natural constraint on the position of the finger so that the errors incurred during the acquisition process are minimized, the identification rate has been greatly improved. Next, the acquired fingerprint images which are of size pixels are transmitted following the serial transmission protocol to the SDRAM of the DSP. After the processes of preprocessing, feature extraction, and matching, the results are shown on a LED where the green light and the red light indicate success and failure of the fingerprint identification task, respectively. Experiments demonstrate that the feasibility of the implemented fingerprint identification system.