基于文档结构特征的碎纸片拼接复原毕业论文模板文档.doc

基于文档结构特征的碎纸片拼接复原 数学与信息科学学院 xxxxx，指导教师： 目 录 摘 要(小四号黑体不加粗) I Abstract t(小四号Times New Roman体加粗) I 1 1 2 问题分析 1 2.1 边缘像素点匹配 1 2.2 文档文字结构控制 2 2.3 纵切碎片的单一横向拼接 2 2.4 横纵切碎片的扩散拼接 2 3 数据预处理 2 4 基于像素点覆盖的匹配度计算方法 3 4.1 横一匹配 3 4.2 封闭匹配 4 4.3 方向场匹配 5 5 基于文档结构的匹配度计算 5 5.1 文档结构控制常量计算 6 5.2 文档结构匹配方法 6 6 碎纸片拼接复原模型的建立 8 6.1 纵切碎片的单一横向拼接 8 6.2.1 建模思路 8 6.2.2 始末碎片 8 6.2.3 模型建立 8 6.2 横纵切碎片的扩散拼接 9 6.2.1 建模思路 9 6.2.2 模型建立 9 7 碎纸片拼接模型求解 9 7.1 纵切碎片的单一横向拼接问题求解 10 7.2 横纵切碎片的扩散拼接模型求解 12 结束语 14 参考文献 15 致谢 17 （注：①目录不加页码； ②中、英文摘要加页码，用罗马数字：Ⅰ，Ⅱ…； ③正文另行加页码，用阿拉伯数字：1，2，3，…．） 摘 要(小四号黑体不加粗)：设计了基于文字特征和文档结构特征的规则碎纸片的高效率拼接复原方法．在问题准备阶段，首先将读为灰度矩阵并经二值化处理将像素点灰度值转化为0-1矩阵根据字高、字宽文结构特征对匹配度进行了校正纵切碎片的拼接复原为寻找有向完全图中哈密顿通路以碎片为节点，经过所有节点一次且仅一次，建立各边权值和最大为目标函数的优化模型横纵切碎片的拼接复原建立求解碎片边缘间最大匹配度的优化模型贪心算法寻求路径关键词：；结；；Abstract t(小四号Times New Roman体加粗)：A efficient splicing restoration method of the regular document scraps is designed based on the text characteristics and structure features．In preparation for the problem，Firstly，the fragments read as gray level matrix then conver the pixel gray value into 0-1 matrix by binarization processing．According to the text characteristics such as strokes continuity and smoothness，design calculation method of the matching degree about pixels coverage．Finally，correct the matching degree based on the structure features such as height，width and line spacing．and get the matching degree between the scraps．As for the splicing recovery problem of longitudinal cut pieces，abstract it as a problem to search Hamiltonian path in directional complete graph，regard the pieces as node，to pass all the nodes one and only one time as constraint conditions，to ensure the sum of edge weights to be maximum as objective function，establish the optimization model; As for the splicing recovery problem of the horizontal and longitudinal cut pieces，abstract it as a cross array problem，set up a optimization model to get maximum matching degree between fragment edges．Next，in order to seek the best path，the diff