手写笔迹的表面法向量提取与鉴别系统  

Surface Normal Vector Extraction and Identification System of Handwriting

作　　者：王执竞 尹凡[1] 陈庆虎[1] WANG Zhi-jing;YIN Fan;CHEN Qing-hu(Electronic Information School,Wuhan University,Wuhan Hubei 430072,China)

机构地区：[1]武汉大学电子信息学院,湖北武汉430072

出　　处：《计算机仿真》2022年第4期156-160,共5页Computer Simulation

摘　　要：针对目前笔迹鉴别中篡改笔迹、临摹笔迹鉴别困难,鉴定人员工作强度大,鉴定设备采购维护成本较高等问题,研制出了一种基于光度立体视觉算法,快速采集光度图片并解算手写笔迹三维信息的应用系统。通过上述系统,可计算得到笔迹表面法向量,还原笔迹凹痕深浅信息,排除背景无深度干扰物的影响。鉴定人员可通过笔迹表面法向量映射的灰度图,分析笔迹书写者的用力习惯。实验结果表明,上述系统对使用人员友好,在帮助鉴定人员判断篡改笔迹内容、临摹笔迹与打印笔迹方面效果明显。In view of the difficulties in the identification of tampered handwriting and copied handwriting, the high work intensity of appraisers and the high cost of procurement and maintenance of identification equipment, an application system based on photometric stereo vision algorithm was developed to quickly collect photometric pictures and solve the three-dimensional information of handwritten handwriting. Through this system, the normal vector of the handwriting surface can be calculated, the depth information of the handwriting dents can be restored, and the influence of the background without depth interference can be eliminated. The appraiser can analyze the handwriting habit of the handwriting writer through the grayscale map mapped by the normal vector of the handwriting surface. Experimental results show that the system is friendly to users, and it is effective in helping appraisers judge the content of tampered handwriting, copy handwriting and print handwriting.

关 键 词：光度立体视觉 法向量 笔迹鉴别 用力习惯 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]