摘要
介绍基于图像识别的自动阅卷系统的设计、核心算法的研究,该系统主要用于批改使用答题卡(供学生使用铅笔涂黑)的试卷。由于采用了检测阈值预估的模式匹配算法,系统较传统的评卷系统相比,对铅笔的质量要求降低,此外评阅速度也比较满意。系统由小型DSP系统和单片图像采集芯片构成,因此体积小,价格也大大低于传统的评卷系统。
An auto- grading system based on image processing is presented. The system can be used to judge the standard test paper. A threshold is calculated before the model matching process,so the system can adapt different pencil and the operation time is very short. So it is a grate improve to traditional test paper judging system. The whole system is mainly composed of DSP and image collection part,which is small and cheap.
出处
《现代电子技术》
2006年第22期119-120,127,共3页
Modern Electronics Technique
基金
军队科研基金资助项目
关键词
图像识别
阚值预估
DSP
图像采集芯片
image identify
threshold calculated
DSP
image collection chip
