以回射器陣列實現光學身分辨識系統

Abstract

由於科技的進步和發展，以往需要大量人力資源的保全管理、倉儲管理系統已逐漸被自動化身分辨識系統(identity recognition system)所取代。目前RFID及1D, 2D條碼正被廣泛地應用著，但這些身分辨識系統仍有許多缺點，如通訊距離短、承載資料量低、安全性差容易被側錄。本研究利用光學回射器陣列設計可調變光學身分識別系統(retro-reflector array optical ID system)來取代現有的RFID及1D、2D條碼系統。其中包含了光學讀取器的建置、兩種身分識別卡的設計與特性研究、身分辨識軟體之資料庫建置與實作。其中讀取器的光路設計與元件配置經改良後已成功將其體積大幅縮小，從原來至少需50x50x20cm的空間縮小為12x12x5cm，使讀取器配置更加微型化、模組化；身分識別卡的部分也透過標準的定量量測機構與對比度分析程式對兩種身分識別卡、三種卡片底材和衰減片之影響作分析與討論，其中調變貓眼製程製作出來的識別卡相較於以調變膜組裝的識別卡讀取角度較小、安全性較高，但回射效率偏低讀取距離較短；識別軟體的部分則是使用傅立葉轉換的方式針對回射器陣列重複性的背景雜訊做過濾，再配合資料庫建置及特徵比對完成應用程式的實作及辨識系統之功能驗證。整體來說本研究所發展的身分識別系統已趨近於實用的階段，但仍有一些環節可以加以改進，相信未來此系統一定能更加實用與方便。

Due to the advance of science and technology, the security management system or warehouse management system, which require a lot of human resources before, are replaced by identity recognition system gradually. Currently, RFID and 1D, 2D barcode are widely utilized, however, these identity recognition system have many shortcomings like limited communication distance, low carrying amount of data, and poor security in avoiding skimming. This research uses optical retro-reflector array to design optical ID system, then, replace existing RFID and 1D, 2D barcode. In the research, the build of optical reader, the design and characters of two kinds of identity card, the build and implementation of identity recognition system’s database. The optical design and the component configuration of the reader have been improved. Their volume is significantly minified from 50x50x20cm to 12x12x5cm, making the deploy becoming more compact and modular. In addition, in the research, through standard quantitative measurements mechanism and the contrast ratio analyze program, we analyze and discuss the influence in two kinds of identity card, three kinds of card substrate, and the ND filters. The identity cards which manufactured through modulation cat’s eye retro-reflector are safer and have narrower read angle than those manufactured through LC modulation film, but at the same time, having shorter reading distance and lower efficiency in retro-reflective. Moreover, the identity recognition system filters repeated background noise of retro-reflective array through fourier transform. Then, accomplish the implementation of application and functional verification of recognition system in concert with database construction and feature matching. Overall, the identity recognition system in the research is close to be applied despite some existing problems. Hope that the system will be more practical and convenient.