Design And Research Of The Ballot Analysis System Based On The Structural Feature Extraction

In contemporary human life, the election has become an essential political activities of every citizen.In the electoral process, people involved in politics through voting. Therefore, the behavior of vote has been becoming one of the democratic state under a wide range of political behavior. Thus,the ballot is the carrier. Ballot is developed by the election statistics units or departments in accordance with national laws, regulations. Voters use them to express their own support, oppose or abstain,then the result will be calculated in accordance with certain procedures and account,which to determine the representative election thought. In the electoral process, we must uphold the principles of fairness and justice, therefore, how to achieve more efficient voting results read and correct statistics is a major issue.From ancient times, in a long period of time, we have using the manual vote count data acquisition. Human solved erverything.so this kind of method is not only cumbersome, resource-consuming, but also the probability of error is relatively large, which contrary to the principles of fairness and justice. Another statistical method is electronic voting and support network, voters elected candidates and electronic statistics,.This approach is relatively simple, but its safety and reliability has long been questioned, and we can not follow up review of evidence, which is not widely used an important reason. The third election voting count is real-time computer control and management system of ballot papers. Real-time ballot system is based on mechanical and electrical integration, software and hardware real-time systems to control the electoral process, in order to achieve the election of intelligent, efficiency, enhancing safety truly. Real-time ballot system is to achieve automatic incoming of the ballot, accurating identification and accounting and automatic storage function of votes. Currently,the real-time system of the computer ballot,with graffiti or other logo increasingly been widely applied,by the using of the computer and the OCR technology.In this paper,we designed a kind of ballot analysis system,which is based on election vote counting system.firstly,the system get the image by using a scanner, rather than using high-definition cameras to collect data, besides.we use TWAIN-related technology and workflow data to get the votes, thus ensuring data integrity and reliability. Nextly, by using the image processing, such as the ballot image preprocessing (graying, binarization, denoising, etc.)and human intervention, we can navigate od the approximate location of the ballot image data, such as the lines of the table position, the location of the code, a tabular frame number data,then we also can locate the accurate identification of the ballot form lines and code,which including the design of hough transform algorithm, the frame of the table location identification, and said custom code identification. Finally, based on structural features,we can extracted handwritten symbol recognition, such as hooks, rings, forks and other defined symbols. In addition,we also design a separate handwritten name entry software,which is to choose another people name’s list entry and statistical work. Finally, with network topology, we design integration of the various parts of the system, which can be server-side statistics and review whether elected.In a word, this ballot analysis system is a new type of intelligent computer processing ballot system,and it achieves automatic ballot entry, automatic identification handwritten symbols, and the results of statistical functions.