摘要
文章通过分析横切、纵切和双面横纵切3种情况,通过MATLAB中神经网络搜寻碎片边缘信息,提取像素值后进行碎片匹配。按照读取图片、搜集数据信息、计算相似度、局部匹配和全局匹配计算拼接后的序列号,必要时进行人工干预的做法,实验表明,该方法有效可行。计算机辅助下的半自动碎片拼接具有广阔的应用前景,能够提高碎片拼接的精确度和鲁棒性。
The paper analyzed cross-cutting,slitting and double-sided fragmentation,searching edge document features by Neural Network of MATLAB,to match pixel values after extraction.Follow this procedure of reading pictures,collecting data,calculating the similarity,partial matching,global matching to get spliced sequence number,doing anual intervention when it is necessary,Experiments showed that the method is effective and feasible.Computer-assisted semi-automatic fragments stitching has broad application prospects,and it can improve the accuracyand robustness of the fragments spliced.
出处
《企业技术开发》
2016年第3期24-26,共3页
Technological Development of Enterprise
关键词
神经网络
碎片拼接
边缘向量
相关系数
neural network
fragments splicing
edge vector
correlation coefficient
