一种基于碎纸片投影特征的拼接算法

提出了一种以影像为基础的碎纸片的拼接重建算法,有别于其他基于形状重建算法,它主要通过扫描和提取碎纸的特征信息,利用纸片水平投影的分布特征,进行关联度计算,创新地将纸片的行信息归结为水平投影向量的形式,降低了纸片本身的复杂度。最后利用图论中的哈米尔顿回路的算法将纸片排序,实现碎纸片的复原工作,并以2013年全国大学生数学建模竞赛B题中的图片为例对中英文两种碎纸片进行了拼接复原实验,实验结果说明该算法有效可行。

A Shredded Document Reconstruction Method Based on Projection Features

In this paper we describe a shredded document reconstruction method. Unlike shape matching method, shredded document are lack of useful shape information. The important data is the distribution of the text and non-text regions for reconstructive algorithms. We propose a novel method to code the distribution of the text and non-text regions to reduce time complexity. We treat the content of documents. Finally, we use the optimum solution in the hamiltonian to sort the order of shreds and complete the document reconstruction, and taking the problem B in 2013 CUMCM as an example. The result shows that the method is effective and feasible.