The Detection and Realization of Data Matrix Code by Accurate Locating

Introduction

As Data Matrix code can express information in both horizontal and vertical directions, it makes full use of the two dimensional plane which increases the density of information greatly, and encodes large data in a small area. At the same time, it has strong error correcting capacity, and the Data Matrix code image can be identified accurately even though it is partially damaged (Tang Li & Liu Fu-Qiang, 2004). Data Matrix code is composed by two color shades, each black or white box of the same size is called a data unit. The data area is surrounded by finder pattern which can be used to determine physical size and location, and the blank area can separate the bar code form its background. Finder pattern is the boundary of its data region which has two adjective sides. The left and lower sides are solid dark lines, and both of them form the “L” boundary which is used primarily to determine physical size, orientation and symbol distortion. The two opposite sides are made up of alternating dark and light modules called railway line. All these signs can be used primarily to define cell structure of the symbol, and can assist in determining physical size and distortion (Leng Biao, 2007). Therefore, Data Matrix code can be composed of four parts: the solid “L” boundary, the opposite dotted “L” boundary of finder pattern, the data region surrounded by the finder pattern and blank area. The concrete structure is shown in Figure 1.

Figure 1.

Structure of data matrix code: (a) Data matrix code; (b) Finder pattern; (c) Data region

Image Preprocessing

For the two-dimensional Data Matrix code automatic identification, due to the physical environment of the image capture device such as cameras and other constraints, noise and various types of distortion are inevitable in the capture of Data Matrix code image (Zou Yan-Xin & Yang Gao-Bo, 2009). In order to improve the reliability of two-dimensional Data Matrix code recognition, more effective image preprocessing method is proposed and the preprocessing steps include graying, filtering, binarization and canny edge detection.

Image Graying

Almost all two-dimensional Data Matrix bar code images acquired through camera and image card are color images. Due to the request of fast and real-time process, the color image should be transformed to the gray image (Wang Jian-Xia, Gao Gui-Long & Yang Hui-Li, 2009). This method is widely used in the field of image processing. Each pixel in color image is determined by tricolor of RGB, and each component has 256 optional values. The method used for transforming is as below:

• 1.

  To calculate the average of three components, and then assign this value to the gray value;

• 2.

  As human eyes are most sensitive to green, different gray-scale image can be obtained by setting different weighting coefficients:

  
  (1)

The transformed result is shown in Figure 2.

Figure 2.

Image graying result