摘要
应用高光谱成像技术识别篡改的字迹,还原原本数据,以及重现同色掩盖的字迹。提取特征波段后,可以对样本数据做出初步的区分。对每个样本依次进行降噪、IsoData、建立眼膜、去除背景和PCA分析处理。并结合数学统计分析提出了识别篡改字迹和重现掩盖字迹的方法。该方法对18种品牌黑色中性笔的篡改字迹识别率和重现掩盖字迹识别率分别为75.8%和57.3%。
Hyperspectral imaging technology is used to recognize handwriting,to restore the original data,and to reappear the samecolor masking handwriting.After extracting the characteristic bands, we can make a preliminary distinction between the sample da-ta.Each sample was then subjected to noise reduction, Iso Data, the establishment of eye mask, removal of background and PCAanalysis.Combined with mathematical statistics analysis, the method of identifying tampering handwriting and reproducing coverhandwriting is put forward.The recognition rate and cover up rate of black neutral pens of 18 brands were 75.8% and 57.3% respec-tively.
作者
王伟利
赵曰峰
WANG Wei-li, ZHAO Yue-feng (School of Physics and Electronic Science, Shandong Normal University, Ji'nan 250014, China)
出处
《电脑知识与技术》
2018年第3期236-237,共2页
Computer Knowledge and Technology
关键词
高光谱成像
字迹
篡改
掩盖
主成分分析
Hyperspectral imaging
Handwriting
Tampering
cover
principal component analysis