复杂背景下条码图像的研究识别.docVIP

复杂背景下二维条码图像的研究识别 目录 摘要 I ABSTRACT II 引言 1 1 绪论 2 1.1 课题研究背景和意义 2 1.2 国内外研究现状 3 1.3 本课题研究内容 4 2 图像的预处理 6 2.1 图像灰度化 6 2.2 图像滤波 7 2.2.1 图像滤波技术 7 2.2.2 高斯滤波 7 2.2.2 中值滤波 8 2.3 二值化阈值选取 9 2.3.1 直方图的峰谷法 9 2.3.2 一维最大熵法 10 2.3.3 Otsu法 11 2.3.4 阈值算法的选取 13 2.4 数学形态学操作 13 2.4.1 腐蚀 13 2.4.2 膨胀 14 2.4.3 开操作 15 2.4.4 闭操作 15 2.5 最大连通分量提取 15 2.6 本章小结 16 3 DM码定位 16 3.1 边缘检测 17 3.1.1 Roberts算子 17 3.1.2 Sobel算子 17 3.1.3 log算子 18 3.1.4 边缘检测算子的选取 18 3.2 Hough变换线段检测 19 3.2.1 Hough变换检测直线原理 19 3.2.2 Matlab 中的Hough变换 20 3.3 图像校正 20 3.3.1 倾角计算 20 3.3.2 图像旋转 22 3.3.3 精确裁剪DM条码 22 3.4 本章小结 23 4 Matlab实验结果分析 24 4.1 DM码识别 24 4.2 DM码解码 31 4.3 实验结果分析 32 4.4 本章小结 33 5 结论与展望 34 致 谢 35 参考文献 36 摘要 Data Matrix二维条码的外观是一个由许多小方格所组成的正方形或长方形符号，其资讯的储存是以浅色与深色方格的排列组合，以二位元码(Binary-code)方式来编码，故电脑可直接读取其资料内容，而不需要如传统一维条码的符号对映表(Character Look-up Table)。Data Matrix二维条码tsu算法；Hough变换；图像旋转 窗体顶端 窗体底端 ABSTRACT The appearance of a two-dimensional bar code is a square or rectangular symbol which made up of many small squares, the information is stored in light and dark colored squares in some way, the computer can read the data directly in the two-dimensional code, without the need for the corresponding table of unified dimensional bar code symbol. However, in the practice of DM code acquisition , because of various factors, the quality of image collected is not as good as expected ,it not only contains the DM code pattern, but also sneaks into a various of backgrounds, so the image of DM code pattern has a complicated background. In view of the above problems, this paper discusses the recognition of 2D barcode in complex background, according to the characteristics of two-dimensional bar code, the solution is as follows :firstly, wipe off the original image noise using Gauss filtering algorithm, binary images via Otsu algorithm, expand and deprive the images using mathematical morphology erosion operation, and then determine the maximum connected region, segment barcode pattern, secondly, test the two longest