基于AI技术的技能操作考核系统  

Skill Operation Examination System Based on AI Technology

作　　者：陈娇英[1] 卞合善 CHEN Jiaoying;BIAN Heshan(Guangxi Vocational and Technical Institute of Industry,Nanning 530001,China;Kebaiwen(Shenzhen)Technology Co.Ltd.,Shenzhen 518101,China)

机构地区：[1]广西工业职业技术学院汽车工程系,广西南宁530001 [2]柯柏文(深圳)科技有限公司,广东深圳518101

出　　处：《北部湾大学学报》2020年第10期80-84,共5页Journal of BeiBu Gulf University

基　　金：2018年广西工业职业技术学院科研项目:人工智能实操考核系统研究(2018036KY001)。

摘　　要：针对当前职业教育技能考核评价弊端,基于AI深度学习算法开发了一套技能操作考核系统。运用SVM算法学习考核标准操作视频;通过多位相机采集考生考核过程的关键动作;采用RNN神经网络构建动作识别模块,比较考生关键动作和标准操作的匹配性;搭建云端评判模块评判考生的考试成绩,实现24 h无人值守的技能操作智能考核。考核学生随到随考,真正体现学分制教育自主择学、择业,高质高效完成学业,又能杜绝教学考核评价不公平现象的出现,真正体现学分制教育改革的内涵及优越性。In view of the disadvantages of vocational education skills assessment and evaluation,a set of skills operation assessment system based on AI deep learning algorithm is developed.In this study,standard operation video was learnt by using SVM algorithm;key actions of examinees'assessment process was collected through multiple cameras;an action recognition module was built based on RNN neural network to compare the matching of candidates'key actions and standard operation;and a cloud evaluation module was built to evaluate candidates'test performance so as to realize the 24-hour unattended skill operation intelligent assessment.With this assessment system,students can take the examination anytime they come.This assessment system can truly reflect the independent choice of study and high-quality and efficient completion of study in the credit system education,and prevent the unfair phenomenon in teaching assessment and evaluation.It can also truly reflect the connotation and superiority of the credit system education reform.

关 键 词：SVM算法 RNN神经网络 智能操作考核 

分 类 号：TP319.[自动化与计算机技术—计算机软件与理论]