The character characteristics of different paper pieces are analyzed in this paper. Several algorithms were deve-loped under reasonable assumptions,which synthesized multi-directional recursion,detection of space between ranks,character pattern recognition and artificial intervention when necessary,and successfully realized the recovery of torn paper pieces. The combination of various algorithms reduced the probability of error splicing effectively and improve the stability of the recovery file scheme. The experimental results after splicing show that the actual splicing effect of the scheme is very satisfactory.%基于不同碎纸片的文字特征进行分析，在合理假设下开发了多种算法，这些算法综合了多向递推、行列间距检测、字符模式识别与必要时的人工干预等多种手段，成功解决了碎纸片的拼接复原问题。同时，多种算法相互结合也有效地降低了错误拼接的概率，提高了复原文件方案的稳定性。拼接完成后的试验结果表明该方案的实际拼接效果非常理想。
展开▼
