标准楷书图像的基本笔划分割---算法设计与实现

论文总字数：29384字

摘 要

目前针对汉字书法的图像处理研究受到了广泛的关注，这种结合了书法艺术与现代图像处理技术的的研究对书法的传承具有重要意义，也对图像处理提出了极高的要求。目前已有的关于汉字图像的处理技术往往是通过数学形态学或者根据汉字骨架进行分解识别，这两项技术往往只能对格式统一印刷体类的书法进行有效识别，对于风格多变的手写书法效果欠佳。

本文研究一种标准楷书笔划分割算法及其软件，作为对书法作品进行识别的前期工作。本文提出一种基于书法字轮廓的笔画分割识别方法。

本文首先分析介绍了当前已有的多种书法图像处理技术的成果和弊端，然后简要介绍了基于书法轮廓进行笔画提取分割的步骤：在对书法图像进行简单的预处理之后，通过寻找拐点分割轮廓段，其后将轮廓段进行重组、配对，然后介绍了本方案使用的三种笔画识别方法：基于斜率识别、底层特征识别和边锋识别。最后分别通过印刷体、书法作品和手写汉字的样本对算法进行检验。发现算法可以对复杂环境下的结构简单的书法进行笔画分解和识别。

关键词：笔画分割、拐点提取、轮廓段配对

ABSTRACT

At present, the research on image processing of Chinese calligraphy has received extensive attention. This combination of calligraphy art and modern image processing technology is of great significance to the inheritance of calligraphy, and it also imposes extremely high requirements on image processing. At present, the processing techniques for Chinese character images are often decomposed and identified by mathematical morphology or according to the Chinese character skeleton. These two techniques can only recognize the calligraphy of the format unified printing type effectively, but the handwritten calligraphy with varied styles powerless.

This paper studies a standard script stroke segmentation algorithm and its software as a preliminary work to identify calligraphy works. This paper proposes a stroke segmentation recognition method based on the outline of calligraphy characters.

This paper first analyzes the results and shortcomings of various existing calligraphy image processing techniques, and then briefly introduces the steps of extracting and segmenting strokes based on calligraphic contours: after simple preprocessing of calligraphic images, the contours are divided by finding inflection points. Then the contour segments are connected and paired. Then the three stroke recognition methods used in this scheme are introduced: based on slope recognition, bottom featurerecognition and edge identification. Finally, the algorithm is tested by the sample of print, calligraphy and handwritten Chinese characters.The discovery algorithm can decompose and identify strokes of simple structures in complex environments.

Keywords: stroke segmentation, inflection point extraction, contour segment matching

目 录

摘要……………………………………………………………………………………………Ⅰ

Abstract………………………………………………………………………………………… Ⅱ

1. 绪论 ……………………………………………………………………………………1

1.1 引言 …………………………………………………………………………………1

1.2 研究问题 ……………………………………………………………………………1

1.3 论文组织 ……………………………………………………………………………3

1.4 程序流程图 …………………………………………………………………………4

1. 预处理 …………………………………………………………………………………5

2.1 归一化 ……………………………………………………………………………5

2.2 二值化 ……………………………………………………………………………6

2.3 去噪平滑 ……………………………………………………………………………7

2.4 本章小结 ……………………………………………………………………………8

1. 笔画提取与识别………………………………………………………………………9

3.1 书法字轮廓提取 …………………………………………………………………9

3.2 提取轮廓拐点………………………………………………………………………10

3.3 独立笔画分析……………………………………………………………………13

3.4 组合笔画分析……………………………………………………………………15

3.4.1复杂轮廓的分割……………………………………………………………15

3.4.2轮廓段的重组………………………………………………………………15

3.4.3轮廓边的配对………………………………………………………………17

3.5笔画识别……………………………………………………………………………19

3.5.1基于斜率的书法字笔画识别………………………………………………19

3.5.2基于底层特征的书法字笔画识别…………………………………………20

3.5.3基于边锋的判定方法………………………………………………………21

3.6本章小结……………………………………………………………………………21

1. 程序测试结果分析……………………………………………………………………23

4.1 印刷体测试分析 …………………………………………………………………23

4.2 手写体测试分析……………………………………………………………………26

4.2.1楷书作品测试分析…………………………………………………………26

4.2.2手写汉字测试分析…………………………………………………………30

4.3 模型优缺点 ……………………………………………………………………33

4.4 本章小结 ………………………………………………………………………33

1. 总结与展望…………………………………………………………………………35

参考文献………………………………………………………………………………………37

附录………………………………………………………………………………………39

剩余内容已隐藏，请支付后下载全文，论文总字数：29384字