摘要
针对现有方法存在切换准确度低的问题,提出基于机器视觉的机器人磨料自动切换方法。先使用CMOS相机采集磨料图像,对图像进行预处理。然后运用灰色关联度来反映磨料的特征信息,计算特征参数点的关联度,通过建立磨料识别优选模型,计算待识别与标准磨料相似性的相对隶属度以提取出磨料特征。最后运用BP神经网络对模型进行训练,将新的磨料图像或数据输入模型中,自动输出分类结果,以控制机器人完成磨料的自动切换。结果表明,应用所提方法,不同种类之间并未存在异常识别现象,10个小组的切换准确程度均为100%,结果符合预期。
Aiming at the problem of low switching accuracy of existing methods,a robot abrasive automatic switching method based on machine vision is proposed.First,use a CMOS camera to capture abrasive images and preprocess the images.Then,the grey correlation degree is used to reflect the characteristic information of abrasives,calculate the correlation degree of feature parameter points,and establish an abrasive recognition optimization model to extract abrasive features by calculating the relative membership degree of similarity between the identified and standard abrasives.Finally,the BP neural network is used to train the model,input new abrasive images or data into the model,and automatically output classification results to control the robot to complete the automatic switching of abrasives.The results showed that there was no abnormal recognition phenomenon among different types using the proposed method,and the switching accuracy of all 10 groups was 100%,which met the expectations.
作者
蔡孟坚
CAI Mengjian(Shanghai Ground and Air Protection Equipment Co.,Ltd.,Shanghai 201602,China)
出处
《自动化应用》
2024年第11期19-21,共3页
Automation Application
关键词
机器视觉
机器人
磨料
自动切换
machine vision
robot
abrasive
automatic switching
