階調濃度調整於圖像化二維條碼整合NFT郵票加值應用

Abstract

郵票作為記錄歷史的有價證券，受到了許多人的收藏，隨著區塊鏈技術發展，非同質化代幣（NFT，Non-fungible Token）也嶄露頭角，使郵票的收藏方式變得更為多元。各國官方郵政陸續推出基於以太坊區塊鏈發行的加密郵票 (Crypto stamp)，加密郵票又以圖像結合嵌入NFT資訊的QR Code（Quick Response Code）作為推廣加密郵票的主要形式。然而，圖像化QR Code在輸出上容易受相關設備影響，導致其輸出後圖像品質不佳，使肉眼辨識及機器讀取條碼資訊變得困難。因此本文將通過不同階調濃度處理之對比，找出符合圖像化QR Code的最佳輸出方式，同時能因應不同類型輸出設備，做出相應調整。本研究採實驗方法，針對不同階調處理的圖像化QR Code，通過訊息嵌入技術和列印輸出後再掃描，每次會將其一種濃度的30張圖像化QR Code結合郵票印刷品通過1200dpi掃描成影像，並以Matlab轉化為600dpi影像，及進行影像分析，計算Module錯誤與Codeword錯誤，通過連續分析不同濃度以30張郵票為一組之平均的辨認錯誤，來比較其是否在QR Code預設之容錯範圍之內，藉此分析出機器可讀取且視覺呈現較佳的影像階調。經實驗與分析後發現影像經階調處理後會影響機器辨識率，在特定階調下可以達到最佳解碼率與良好的視覺樣貌，且黑白圖像與彩色圖像比較中，顯示彩色圖像的整體錯誤率較低，並將所製作的城市意象QR Code NFT郵票進行加值應用。Stamps, historically valued as securities, have been widely collected. With the advent of blockchain technology, Non-fungible Tokens (NFTs) have introduced new dimensions to stamp collecting. National postal services are now issuing crypto stamps on the Ethereum blockchain, primarily using image-embedded QR codes containing NFT data. However, the quality of these visual QR codes can be compromised by output device limitations, making them difficult to read both visually and mechanically.This study aims to identify the optimal output method for visual QR codes by comparing different tonal treatments and adjusting for various output devices. An experimental approach was used, where 30 samples of QR codes with different tonal treatments were printed, scanned at 1200 dpi, and analyzed in Matlab at 600 dpi. The study calculates module and codeword errors to determine if the tonal variations fall within the QR code's error tolerance range, thus identifying the most machine-readable and visually appealing tonal quality. Results indicate that tonal processing affects machine recognition rates, with specific tonal settings yielding the best decoding accuracy and visual quality. A comparison between monochrome and color images revealed that color images have a lower overall error rate. The findings were applied to enhance city-themed QR code NFT stamps.