Regular Text Fragmentation Based on Pixel Points

For a page that is both vertical and horizontal rule Chinese fragments and a page of regular double-sided printed English document fragments, this paper respectively proposes an image matching algorithm based on line and column and a two-dimensional image link matching algorithm. Firstly, the edge pixel points are obtained through fragmentation, and the obtained pixel points are converted into corresponding gray-scale values. Then, the gray value of each fragment was matched, and the matching degree was taken as the basis for the splicing between fragments. At last, through the Matlab software, the sequence number of pieces of Mosaic can be calculated and output by using the corresponding pieces stitching algorithms. The experimental results show that the algorithm can obtain the result of fragment Mosaic correctly and quickly, which greatly improves the efficiency of reassembly.