中密度纤维板表面缺陷实时检测系统  

Design of detection system for MDF surface defects

作　　者：高鹏威 高远 杨建华 吴健[1] 闫磊[1] GAO Peng-wei;GAO Yuan;YANG Jian-hua;WU Jian;YAN Lei(Beijing Forestry University,Beijing 100083,China;Beijing Institute of Forestry Machinery,National Forestry and Grassland Administration,Beijing 100029,China;Forestry New Technology Research Institute,Chinese Academy of Forestry,Beijing 100091,China)

机构地区：[1]北京林业大学,北京100083 [2]国家林业和草原局北京林业机械研究所,北京100029 [3]中国林业科学研究院林业新技术研究所,北京100091

出　　处：《林业和草原机械》2020年第1期47-50,共4页Forestry and Grassland Machinery

基　　金：国家重点研发计划项目:木基材料与制品增值加工技术(2018YFD0600304);中央级公益性科研院所基本科研业务费专项资金:中密度纤维板外观质量检测关键设备研究(CAFYBB2017MB033)。

摘　　要：中密度板加工生产过程中,表面缺陷检测环节现有方法仍然以人工为主,检测速度慢、效率低。针对中密度纤维板的油污、松软、黑斑、杂物等常见的6类缺陷,设计了一种基于线阵相机的中密度纤维板表面缺陷实时检测系统。提出了一种改进的CenterMask算法,该算法对原有特征提取网络进行了网络改进。在中密度板表面缺陷数据集中的测试结果表明:在检测速度大于6FPS的算法中,改进的CenterMask算法对有着最高的缺陷分割准确率。In the process of MDF processing and production,the existing technology of surface defect detection is still dominated by manpower,and the detection speed is slow and the efficiency is low.In this paper,a real-time detection system of MDF surface defects based on linear array camera is designed for six kinds of common defects,such as oil,soft,black spots and sundries.This paper proposes an improved Center Mask algorithm,which improves the original feature extraction network.The test results in the MDF surface defect data set show that the improved Center Mask algorithm has the highest defect segmentation accuracy in the algorithm with detection speed greater than that of 6fps.

关 键 词：中密度板 表面缺陷检测 线阵相机 改进的CenterMask 

分 类 号：TS653[轻工技术与工程] TP391.41[自动化与计算机技术—计算机应用技术]