一個照相手機上的牛眼碼解碼系統

Abstract

在本論文中，我們提出了一個牛眼碼(MaxiCode)的解碼系統，並實作於照相手機上。牛眼碼是一種尺寸固定的矩陣式二維條碼，主要為運輸業所使用來儲存貨物運送資訊。論文中，我們利用照相手機對牛眼碼取像，再對拍下的影像作處理。首先，我們利用邊緣點的資訊，對影像進行二值化的處理。對二值化後的影像，利用細線化處理後的結果，尋找最大的連通分量(Connected Components)，以定位牛眼碼的探測圖案(Finder Pattern)。對於牛眼碼在圖片中可能旋轉的情形，利用定位圖案(Orientation Patterns)將條碼轉正。而拍照所造成條碼邊界的形變，利用反透視轉換(Inverse Perspective Transformation)來校正。最後以水平投影的方式擷取出資料點的長度，轉換成對應的模組來進行解碼的動作。實驗結果顯示本系統能正確且有效率的對牛眼碼圖片進行解碼。

In this thesis, we construct a MaxiCode decoding system, and we implement it on the camera mobile phone. MaxiCode is a fixed-size matrix 2D barcode, and it is often used by transportation industry to record the transportation information of packages. In our system, we process the MaxiCode’s picture which is taken by the camera phone. Firstly, the edge points of the picture are used for the image binarization. Next, we use the thinned connected components to locate the Finder Pattern of MaxiCode, and we find the Orientation Pattern of Maxicode to rotate the code into the right position. Then we extract the boarders of MaxiCode, and we use inverse perspective transformation to rectify the code shape. Finally, the run-length of data points are extracted by horizontal projection, and the modules corresponding to the run-length are taken to decode. Experimental results show that the system can decode the MaxiCode image correctly and efficiently.