摘要
为了研究计算机图像识别技术在皮革外观缺陷检测中的应用,通过2600X反射式激光光学电镜和计算机学习算法相配合,可以代替部分人工视觉。通过搭建机器检测平台,对不同缺陷的皮革图像采集成册,进行空间卷积处理、模糊皮革纹理背景,把皮革缺陷中的鞣制缺陷、染色缺陷、材质缺陷呈现出来,完成检测分类的任务。采用人工智能算法对鞣制缺陷、染色缺陷、材质缺陷的敏感度和特异度进行对比,论证得出人工智能算法作为一种重要的非接触式测量方法,可以提供较高的检测精度,对提高产品质量和生产效率,提升企业竞争力都有重要的现实意义。
In order to study the application of computer image recognition technology in leather appearance defect detection,the combination of 2600X reflective laser optical microscope and computer learning algorithm can replace some artificial vision By building a machine detection platform the leather images of different defects were collected into volumes,spatial convolution processing was carried out,the leather texture background was blurred,and the tanning defects,dyeing defects and material defects in leather defects were presented to complete the task of detection and classification The sensitivity and specificity of tanning defects,dyeing defects and material defects were compared by using artificial intelligence algorithm.It is demonstrated that artificial intelligence algorithm,as an important non-contact measurement method,can provide high detection accuracy,and has important practical significance for improving product quality,production efficiency and enterprise competitiveness.
作者
张娜娜
ZHANG Na-na(Shanghai Sipo Polytechnic,Shanghai 201300,China)
出处
《中国皮革》
CAS
2022年第7期20-23,共4页
China Leather
基金
上海市职业教育协会一般课题(YB202009)
上海思博职业技术学院教改课题(SPJG1905)
上海市高等教育学会课题(Y1-70)。
关键词
图像识别
外观缺陷
光学电镜
空间卷积
人工智能
image recognition
appearance defect
optical electron microscope
spatial convolution
artificial intelligence
