一种基于灰度的碎纸片拼接算法.

In this paper, an approach of recovery of paper scraps was proposed. First of all, adjacent bits of paper text line feature were selected, and each piece of paper was used as a node of the graph. It established a grid graph model based on the line feature of text, and narrowed the set of matching scraps of paper by considering both sides of pieces line the matching. Function values of both positive and negative similarity were added as similarity of scraps of paper, when the similarity function value reached minimum. It shows the difference of gray value of two scraps of paper writing disconnection can reach the minimum. Finally, the simulated annealing mechanism was introduced to optimize the matching process of the set of matching pieces of paper. The results showed that the algorithm can solve the problem of more complex scraps of splicing. [ABSTRACT FROM AUTHOR]