【摘要】：隨著計算機信息技術(shù)的發(fā)展,網(wǎng)絡信息化技術(shù)在考試中的應用規(guī)模也越來越宏大�？荚囀墙虒W的重要環(huán)節(jié),考試數(shù)量的增多和考試規(guī)模的擴大,傳統(tǒng)考試內(nèi)容限定的選擇題和問答題逐漸演變?yōu)榧瘑芜x題、多選題、客觀主觀選做題、主觀性試題、發(fā)散性思維題等題型為一體的多元化發(fā)展態(tài)勢,傳統(tǒng)閱卷方式對評卷誤差的把握、閱卷后的分數(shù)統(tǒng)計分析,試卷的運送存儲等都無法做到精確。網(wǎng)絡閱卷系統(tǒng)是以提高閱卷速度,減小主觀題閱卷誤差,便于分數(shù)的統(tǒng)計分析及試卷的管理存檔為目的,以考試的公平公正性原則為宗旨的新型閱卷方式。本文采用“互聯(lián)網(wǎng)+”的設計模式、數(shù)字圖像處理技術(shù)以及大數(shù)據(jù)分析技術(shù),將紛繁復雜的閱卷工作和錯綜復雜的數(shù)據(jù)統(tǒng)計分析交由系統(tǒng)的“掃描和評閱”自動完成。系統(tǒng)主要分客觀題閱卷和主觀題閱卷兩大部分,試卷經(jīng)過圖像掃描,系統(tǒng)自動識別考生基本信息,客觀題由系統(tǒng)自動判分,主觀題由系統(tǒng)隨機將分割后的題塊分發(fā)到計算機終端,由閱卷教師進行網(wǎng)絡閱卷,閱卷過程中注入多評機制,保證閱卷公平性,系統(tǒng)自動進行成績統(tǒng)計和校驗。網(wǎng)絡閱卷將計算機網(wǎng)絡技術(shù)與教師閱卷經(jīng)驗相結(jié)合,既減少了閱卷錯誤又相應地提高了閱卷效率。本文在網(wǎng)絡閱卷整個系統(tǒng)的設計實現(xiàn)中,主要做了如下工作:(1)在閱讀相關文獻的基礎上分析對比國內(nèi)外研究背景和現(xiàn)狀,重點分析國內(nèi)幾家系統(tǒng)的優(yōu)缺點,闡述了迫切開發(fā)一款效率高、功能強的網(wǎng)絡閱卷系統(tǒng)的重要性,針對用戶真實需求,結(jié)合網(wǎng)絡應用于教育的教育技術(shù)學理論,對整個系統(tǒng)的開發(fā)方案進行分析與設計;(2)分析了圖像預處理的相關技術(shù),針對幾種常用的圖像預處理方法,找出適應于本系統(tǒng)的圖像預處理技術(shù),對電子掃描答題卡作圖像預處理操作,實現(xiàn)了客觀題答案識別。最后將卷積神經(jīng)網(wǎng)絡應用到手寫答案和考號識別領域,實現(xiàn)對涂卡客觀題答案和手寫體答案及考號的自動識別,有效地提高了系統(tǒng)在識別方面的準確率和效率。(3)最后本系統(tǒng)采用B/S開發(fā)模式,利用SSH框架Java技術(shù)和MySQL數(shù)據(jù)庫,最終開發(fā)實現(xiàn)了整個閱卷系統(tǒng)。
[Abstract]:With the development of computer information technology, the application scale of network information technology in examination is more and more large. Examination is an important link in teaching. With the increase of the number of examinations and the expansion of examination scale, the traditional examination content limited multiple choice questions and questions gradually evolved into single topic selection, multi-topic selection, objective subjective selection of questions, subjective test questions. Divergent thinking questions and other questions as one of the diversified development situation, the traditional marking of the error of the paper evaluation, the statistical analysis of scores after marking, the transport and storage of test papers can not achieve accuracy. The network marking system is a new type of marking method which aims at improving the marking speed, reducing the error of the subjective questions marking, facilitating the statistical analysis of the scores and the management and archiving of the examination papers, and taking the principle of fairness and fairness in the examination as the purpose. In this paper, the design mode of "Internet", digital image processing technology and big data analysis technology are adopted, and the complicated marking work and the complicated data statistical analysis are automatically completed by the system "scan and review". The system is mainly divided into two parts: objective question marking and subjective question marking. After image scanning, the system automatically recognizes the basic information of the examinee, and the objective questions are automatically graded by the system. The subjective questions are distributed randomly to the computer terminal by the system, and the network marking is carried out by the marking teachers. The multiple evaluation mechanism is injected into the marking process to ensure the fairness of the marking, and the system automatically carries out the results statistics and verification. The combination of computer network technology and teachers' marking experience not only reduces the error of marking but also improves the efficiency of marking. The main work of this paper is as follows: (1) on the basis of reading related literature, this paper analyzes and compares the domestic and foreign research background and current situation, and analyzes the advantages and disadvantages of several domestic systems. This paper expounds the importance of developing a network marking system with high efficiency and strong function. According to the real needs of users, it combines the educational technology theory of network application in education. Analysis and design of the whole system development scheme; (2) analysis of the image preprocessing technology, aiming at several commonly used image preprocessing methods, find out the image preprocessing technology suitable for this system. The electronic scan answer card is preprocessed and the objective answer recognition is realized. Finally, the convolution neural network is applied to the field of handwritten answer and test number recognition, which realizes the automatic recognition of the objective answer and handwritten answer and the test number. The accuracy and efficiency of the system in recognition are improved effectively. (3) at last, the system adopts the B / S development mode, uses SSH framework Java technology and MySQL database, and finally develops and realizes the whole marking system.
【學位授予單位】：石家莊鐵道大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：G434

【參考文獻】

相關期刊論文 前10條

1 曾雪瓊;黎杰;;基于卷積神經(jīng)網(wǎng)絡的時頻圖像識別研究[J];機械與電子;2016年05期

2 胡青;劉本永;;基于卷積神經(jīng)網(wǎng)絡分類的說話人識別算法[J];信息網(wǎng)絡安全;2016年04期

3 陽哲;;卷積神經(jīng)網(wǎng)絡在印章編號識別中的應用[J];現(xiàn)代計算機(專業(yè)版);2016年04期

4 宋波;王啟春;;卷積神經(jīng)網(wǎng)絡在路牌識別中的應用[J];公路交通技術(shù);2015年05期

5 王振;高茂庭;;基于卷積神經(jīng)網(wǎng)絡的圖像識別算法設計與實現(xiàn)[J];現(xiàn)代計算機(專業(yè)版);2015年20期

6 蔡娟;蔡堅勇;廖曉東;黃海濤;丁僑俊;;基于卷積神經(jīng)網(wǎng)絡的手勢識別初探[J];計算機系統(tǒng)應用;2015年04期

7 余永維;殷國富;殷鷹;杜柳青;;基于深度學習網(wǎng)絡的射線圖像缺陷識別方法[J];儀器儀表學報;2014年09期

8 鄭胤;陳權(quán)崎;章毓晉;;深度學習及其在目標和行為識別中的新進展[J];中國圖象圖形學報;2014年02期

9 徐虹;高偉;;美國大學考試形式及其對我國考試評價改革的啟示[J];教育探索;2013年07期

10 徐姍姍;劉應安;徐f;;基于卷積神經(jīng)網(wǎng)絡的木材缺陷識別[J];山東大學學報(工學版);2013年02期

相關博士學位論文 前1條

1 鄭澤忠;基于高分辨率航空影像高速公路汽車目標檢測算法研究[D];西南交通大學;2010年

相關碩士學位論文 前8條

1 常歡;基于卷積神經(jīng)網(wǎng)絡的孤立手寫體漢字識別研究[D];安徽大學;2015年

2 李敏;基于BS模式的網(wǎng)絡閱卷系統(tǒng)的設計與實現(xiàn)[D];吉林大學;2014年

3 張愛娟;基于結(jié)構(gòu)特征點的字符分割技術(shù)的研究[D];西安電子科技大學;2014年

4 陳寅;植物葉形狀與葉脈結(jié)構(gòu)的自動分類研究[D];浙江理工大學;2013年

5 謝琪林;網(wǎng)上閱卷系統(tǒng)設計與實現(xiàn)[D];電子科技大學;2011年

6 張智軼;基于圖像處理的燃燒物識別技術(shù)的研究[D];東華大學;2011年

7 孫偉;基于FPGA的圖像分割提取系統(tǒng)研究[D];南京航空航天大學;2010年

8 王添翼;基于卷積網(wǎng)絡的三維模型特征提取[D];吉林大學;2006年

，

本文編號：2265253