基于机器视觉的答题卡自动评分系统设计与实现  

Design and implementation of automatic scoring system for answer sheet based on machine vision

作　　者：高强 潘俊[2] 刘薇[1] 高文丽 Gao Qiang;Pan Jun;Liu Wei;Gao Wenli(Guangzhou Railway Polytechnic,Guangzhou,Guangdong,511300,China;Guangzhou Civil Aviation College;Shangrao Children's Normal College)

机构地区：[1]广州铁路职业技术学院,广东广州511300 [2]广州民航职业技术学院 [3]上饶幼儿师范高等专科学校

出　　处：《计算机时代》2023年第12期180-183,共4页Computer Era

基　　金：2022年广东省普通高校特色创新项目“基于机器视觉的飞行签派员疲劳检测关键技术研究”(2022KTSCX227);2022年广东省高等教育学会课题“人工智能技术在高校教学阅卷中的应用与实践研究”(22GYB013)。

摘　　要：为了降低光标阅读机使用成本与实现答题卡自动化识别,设计并实现了一种基于机器视觉的答题卡自动识别系统。首先对答题卡图像进行滤波、灰度变换、二值化等预处理操作;然后采用Hough变换实现双直线检测,基于该直线对答题卡进行矫正及区域分割;最后通过填涂区域的几何属性及像素点值实现选项识别判断。测试结果表明,系统界面友好、灵活、使用方便、成本低廉,答题卡识别准确率、效率高,具有一定的应用价值。In order to reduce the cost of using cursor readers and realize automatic identification of answer sheets,an automatic identification system of answer sheets based on machine vision is designed and realized.Firstly,preprocessing operations such as filtering,grayscale transformation and binarization are performed on the answer sheet image.Then the Hough transform is used to achieve double line detection,based on which the answer sheet is corrected and segmented.Finally,option identification and judgment are achieved through the geometric attributes and pixel values of the filled area.The test results show that this system has a friendly interface,flexibility,easy to use,low cost,high accuracy and efficiency in answering sheet identification,and has certain application value.

关 键 词：机器视觉 答题卡 图像识别 评分 

分 类 号：TP311.25[自动化与计算机技术—计算机软件与理论]