社区选举系统选票中的表格识别算法  

Form recognition algorithm in ballot of community election system

作　　者：邝振 崔喆[1,2] 

机构地区：[1]中国科学院成都计算机应用研究所,成都610041 [2]中国科学院大学,北京100049

出　　处：《计算机应用》2017年第A02期179-182,共4页journal of Computer Applications

基　　金：四川省科技支撑计划项目(2015GZ0088);"西部之光"联合学者项目

摘　　要：针对社区选举系统识别选票图像中的表格采用的同步道方法,造成选票设计复杂,需在系统中定义较多关于信息块位置信息的参数,和在印刷过程中出现表格位置偏差无法识别选票的问题,提出了先利用投影法确定表格中横纵直线的坐标,然后构造表格特征点矩阵来描述表格结构的表格识别算法。实验结果表明,与采用同步道识别表格的方法相比,只需要在系统中指定信息块的相对位置信息,该算法即可通过识别出的选票表格中横纵直线坐标和构造的表格特征点矩阵准确并方便地提取出信息块的图像进行下一步的识别,简化了选票设计,降低了对选票印刷的要求,对表格位置出现偏差的选票也能够进行正常识别,进一步提升了选举系统的可操作性和容错性。Concerning the algorithm for identifying the tables in the ballot image based on recognition of synchronization tracks, it is complicated to design the ballot, too many parameters of the information block position information need to be defined, the ballot may be recognized inaccurately because of the table position deviation during the printing process in the community election system. A ballot recognition method was proposed to use projection method to locate the coordinate of the vertical and horizontal lines in the table first, and then the table feature point matrix was constructed to describe the form structure. The experimental results showed that compared with the method of using the track recognition table, only the relative position information of the information blocks needed to be specified in the system, and by using the algorithm and the information blocks of the image could be accurately and easily extracted through the identification of the ballot sheet in the vertical and horizontal coordinates and the table feature point matrix, the design of the ballotcould be simplified, the requirement for the printing of the ballot could be reduced, and the ballot with the deviation of the position of the table could also be effectively identified. The algorithm further improves the operability and fault tolerance of the election system.

关 键 词：表格识别 选举系统 投影法 表格结构 表格特征点矩阵 

分 类 号：TP311.132[自动化与计算机技术—计算机软件与理论]