人机协作中安全手套佩戴检测的模式识别研究

为了更有效、智能地解决人机协作中的安全问题,以人手安全作为人机协作安全的代表,提出了一种基于模式识别技术的人机协作人手安全检测模型。人机协作中,安全手套对于人手保护不可或缺。模型首先通过机器对人手图像获取、人手特征提取形成人手状态训练集,再利用支持向量机训练形成人手安全分类器,以达到智能地判断人机协作中人手是否戴安全手套。选取人机协作现场照片为样本源,采集工人不安全行为场景数据,验证应用基于模式识别的方法智能检测人机协作中工人佩戴是否安全手套的可行性。实例验证结果表明,人手安全分类器可以有效地判断人手的安全状态并能及时控制机器,准确率达到96. 12%。

Research on Pattern Recognition of Safety Gloves Wear Testing in Human-machine Cooperation

To solve the safety problem of human-machine cooperation more effectively and intelligently,a human-machine cooperation safety model based on pattern recognition technology is proposed based on human hand safety as a representative of human-machine cooperation safety check.In human-machine cooperation,safety gloves for hand protection is indispensable.In the model,firstly the hand state training set is formed based on the hands of image acquisition and human hand feature extraction through the machine and then the support vector machine is applied to train and form hand safety classifier,to intelligently judge in cooperation hand whether it wears safety gloves.Human machine collaboration scene photos are selected as sample sources and workers’unsafe behavior scene data is collected to verify the feasibility whether workers wear safety gloves in human-machine cooperation in applying intelligent recognition method based on pattern recognition detection.The example verification shows that the human hand safety classifier can effectively judge the safety state of the human hand and control the machine in time,and the accuracy rate is 96.12%.