摘要
胶黏剂可将小尺寸木质材料单元黏接成人造板、胶合板和胶合木等形式,应用于木结构建筑、承重墙、装饰、地板和桥梁等工程材料中.胶合性能是评价胶黏剂胶黏能力的重要依据,与其相关的指标主要包括胶合强度、施胶分布、渗透性、剪切强度、剥离率、木破率等.近年来,机器视觉技术已被广泛用于辅助评价胶合性能.为更详细、全面了解基于机器视觉的胶合性能测量技术,本研究在收集国内外胶合性能测量技术相关文献的基础上,概述基于机器视觉的图像采集技术、图像预处理技术、图像分割技术和图像识别分类,阐述机器视觉技术在胶黏剂分布、渗透性和木破率3方面的应用,讨论基于机器视觉技术无法准确分割数字图像导致测量精度低的相关问题,总结机器视觉技术胶合性能测量技术优化方法,并提出今后研究展望:1)实现胶接接头上胶黏剂与木质材料的精准识别,特别是浅色胶黏剂与基材的识别,尤其包括刨花板和纤维板等板材的胶点与细小的木质单元;2)深入探究不同测量对象的像素级单位特征与机器视觉的测量机制;3)基于机器视觉技术进行科学、健全和体系的评价机制研究;4)深入探究图像采集的光照条件、采集设备、样品放置等环境因素对采集图像质量的影响机制;5)建立胶合性能评价数据库,为深度学习或胶合性能深入分析提供数据来源.
Abstract
Adhesives can make small-sized wood material unit into wood-based panels,plywood,and glued wood forms,which are used in engineering materials such as wood-frame buildings,load-bearing walls,decorative flooring,and bridges.The important basis for evaluating the gluing ability of adhesives is the gluing performance,which mainly includes the distribution of glue application,penetration ability,shear strength test,peeling test,wood breakage percentage measurement,and other indicators.With the popularization of machine vision technology,it has been applied as an important technique to assist in the evaluation of gluing performance.In order to have a more comprehensive understanding of the gluing performance measurement technology of machine vision,this paper outlines the image acquisition technology,image pre-processing technology,image segmentation technology,and calculation methods of machine vision technology on the basis of collecting domestic and foreign gluing performance measurement technology research literature.Secondly,it discusses the application of machine vision technology in the three aspects of adhesive sizing,adhesive permeability,and wood breakage percentage of glued surfaces.At the same time,the problem that machine vision technology can not segment digital image accurately,resulting in poor measurement accuracy is discussed.Finally,the optimization method of adhesive property measurement technology based on machine vision technology is summarized,and future research prospects are proposed:1)to achieve accurate identification of adhesives and wood materials on adhesive joints,especially the identification of light-colored adhesives and substrates,especially the glue points and small wood units of particleboard and fiberboard;2)Deeply explore the pixel-level unit characteristics of different measurement objects and the measurement mechanism of machine vision;3)Research on scientific,sound and systematic evaluation mechanism based on machine vision technology;4)Explore the influence mechanism of environmental factors such as illumination conditions,acquisition equipment and sample placement on image quality;5)Establish a gluing performance evaluation database to provide data sources for deep learning or in-depth analysis of gluing performance.
国家科技期刊平台
NSTL
万方数据