碎纸片拼接复原模型

Scraps of Paper Splicing Recovery Model

摘要: 本文的主要目的是探究怎样对边缘规则的碎纸片进行快速拼接复原。首先针对以先横再纵的方式切割得到边缘整齐的碎纸图片，使用Matlab软件图像处理功能把图像转化为相应的数值矩阵，再对得到的数值矩阵元素进行二值化处理得到0-1矩阵。应用聚类分析法将具有相同位置特性的碎纸片归类为一组，再将某个图片像素点矩阵中的第一列和最后一列与其他图片像素点矩阵的第一列和最后一列进行边缘匹配，根据边缘匹配度来确定复原图片顺序，再根据边缘匹配度将他们连接、检查并做人工干预可得拼接后的横行纸片，最后将横行纸片的长边用同样的方法做边缘匹配可将行与行之间拼接起来，再做人工调整得到最优结果。

Abstract: The main purpose of this paper is to explore how the paper scraps of neat edges could be restored. First, we can obtain the paper scraps of neat edges after cross cut and then length cut. With the help of the image-processing software, Matlab, we can transfer the images into corresponding matrix, and then binarize the matrix and get the 0-1 matrix. We classify the paper scraps with the same location features using cluster analysis. Then we match the first column and the last column in certain pixel matrix with the corresponding columns in other pixel matrix based on the edge matching. According to the edge matching, we determine the order of the pictures, splice them and then check them with necessary manual adjustment, so we get the transverse paper. Finally, the same method can be applied to the transverse paper while we match the lines instead of the columns and then we get the restored paper after manual adjustment.