基于高斯-马尔可夫随机场的木材表面缺陷类型识别

为了识别死节、活节和虫眼三种木材表面缺陷类型,本文采用高斯-马尔可夫随机场模型提取木材表面缺陷图像的纹理参数,结合缺陷区域的矩形度和伸长度两个几何特征,形成14维特征向量.设计三层BP神经网络来识别缺陷的类型.试验表明,三种缺陷的整体识别正确率达到96.67％,验证了该方法的有效性.     

木材缺陷识别方法的研究

主要研究运用数字图像处理技术对木材表面缺陷的纹理特征进行具体研究的理论方法和可行性,并通过分析比较,探索一个分形理论和小波结合的的处理方案,完善和丰富木材科学理论,提高木材的利用率。     

Soft measurement of wood defects based on LDA feature fusion and compressed sensor images

We proposed a detection method for wood defects based on linear discriminant analysis(LDA) and the use of compressed sensor images. Wood surface images were captured, using a camera Oscar F810C IRF camera,and then the image segmentation was performed, and the defect features were extracted from wood board images. To reduce the processing time, LDA algorithm was used to integrate these features and reduce their dimensions. Features after fusion were used to construct a data dictionary and a compressed sensor was designed to recognize the wood defects types. Of the three major defect types, 50 images live knots, dead knots, and cracks were used to test the effects of this method. The average time for feature fusion and classification was 0.446 ms with the classification accuracy of 94%.     

Wood defects during industrial-scale production of thermally modified Norway spruce and Scots pine

This research investigates wood defects, particularly the formation of surface cracks, during the production of thermally modified wood and its exposure to cyclic moisture changes. Boards of Norway spruce and Scots pine originating from different steps within the production of ThermoWood® were collected and wood defects were investigated at macroscopic and microscopic scale. Subsequently, the wood was exposed to capillary wetting cycles to record its sensitivity towards cracking. After the modification process, typical anatomical defects of conventional kiln-drying became more frequent and severe, with the magnitude being to some extent depending on the presence of defects in the raw material. At microscopic scale, damages to ray parenchyma and epithelial cells as well as longitudinal cracks within the cell walls of earlywood tracheids were evident in thermally modified wood. Despite a lower water uptake and higher dimensional stability, thermally modified wood was more sensitive to surface cracking during wetting cycles than unmodified wood, i.e. at the outside face of outer boards (near bark). For limiting surface cracking of thermally modified wood during service life, the use of high-quality raw material, the exposure of the inside face of the boards (near pith) and the application of a surface coating are considered beneficial.     

Feature extraction of wood-hole defects using wavelet-based ultrasonic testing

The primary bottleneck to extracting wood defects during ultrasonic testing is the accuracy of identifying the wood defects. The wavelet energy moment was used to extract defect features of artificial wood holes drilled into 120 elm samples that differed in the number of holes to verify the validity of the method. Wavelet energy moment can reflect the distribution of energy along the time axis and the amount of energy in each frequency band, which can effectively extract the energy distribution characteristics of signals in each frequency band; therefore, wavelet energy moment can replace the wavelet frequency band energy and constitute wood defect feature vectors. A principal component analysis was used to normalize and reduce the dimension of the feature vectors. A total of 16 principal component features were then obtained, which can effectively extract the defect features of the different number of holes in the elm samples.     

基于Faster R-CNN的实木板材缺陷检测识别系统

我国木材资源有限,为了提高木材的利用率,采用机器视觉来实现木材缺陷快速而稳定的检测,不仅可以克服人工检测的低效率和木材缺陷识别的低准确率,而且对提高木材加工企业的智能化水平具有重要意义。为了高效、快速、准确地进行无损检测,采用深度学习方法,建立了一种基于快速深度神经网络的实木板材缺陷识别模型。首先采用Resnet V2结构对采集到的实木板材缺陷图像进行特征提取,然后应用该模型对节子、孔洞等实木板材缺陷进行训练学习,最后构建了Faster R-CNN检测框架,并使用tensorflow开发平台对节子、孔洞等实木板材缺陷进行预测输出。具体选取了2 000块杉木样本,通过旋转对原始的实木板材图像进行数据扩充,扩充后图像的80%作为训练集,20%作为验证集来进行仿真。仿真结果表明,该模型对实木板材节子缺陷检测正确率为98%,对实木板材孔洞缺陷检测正确率为95%,验证了将深度学习算法应用于实木板材缺陷检测中的有效性。     

基于高斯-马尔可夫随机场的板材表面纹理分析

阐述了高斯-马尔可夫随机场模型的基本原理,并针对板材表面纹理建立了高斯-马尔可夫随机场模型,以利用其对板材表面纹理进行分析;详细的研究了高斯-马尔可夫随机场模型的参数估计过程,在试验数据的基础上,验证了高斯-马尔可夫随机场模型在纹理分析中应用的可行性。     

基于融合彩色特征木材表面颜色分类的研究

针对木材表面颜色自动分类的难题,在RGB颜色空间,将R、G、B三个颜色矩阵融合成一个特征矩阵,再对这个特征矩阵提取颜色三阶矩参数作为木材表面颜色分类的特征参数,设计了适合木材表面颜色分类的BP神经网络分类器,分类识别率达到98.67%,验证了提取特征参数的有效性。     

A MACHINE VISION SYSTEM FOR INSPECTING WOOD SURFACE DEFECTS BY USING NEURAL NETWORK

m0DUrn0NStUapngthemachinevislonsystemistOraiseautomahonlevelOfdefectsinspechonandtOimProvePredctsquality.Inthispaper,themainfunhonOfthisknOwledge-basedmachinevisionsystemistodetectUndsirable"defecs"thatcanamparonthesutheeoftheroughwoodlumbers.F0ramachinevisionsystemtohandieacomPlexProblem,alotofaPriorknOwedgemustbeembeddedideitinanorsanibomanner.Itisimportanttokeepdomain-specificknOWedgeseparatefrOmageneralknOwedgeaboutthenetWrkinSpechonPrObIeminordertOmakeasystemeasilyadaPtabletodi…     

机器视觉在木材缺陷检测领域应用研究进展

为了提高木材利用率，采用机器视觉来实现木材缺陷快速而稳定的检测，不仅可以克服人工检测效率低、劳动强度大、准确率低等弊端，而且对提高木材加工企业的智能化水平具有重要意义。文中概述了机器视觉检测技术及设备的国内外研究现状，介绍基于机器视觉检测木材缺陷所涉及的相关理论和算法研究，以及相关图像处理算法的优缺点；针对机器视觉应用在木材缺陷检测领域存在的不足，提出机器视觉木材表面检测应进一步向人工智能方向发展，以提高木材缺陷检测效率及准确性。     

木材等离子体改性研究进展

综述了木材等离子体改性研究进展, 包括等离子体表面改性技术、木材亲水性和疏水性、木材表面形貌和表面化学组成、木材胶接性、木材阻燃性、竹材等离子体改性。     

黄毛青冈等四种硬木干燥特性的研究

对黄毛青冈、高山栲、滇石栎和栓皮栎等四种硬阔叶树材的构造、密度、全干缩率、弦、径向干缩变化、用百度法和气干等进行了观察和试验。结果表明四种木材导管内含有丰富或较多的侵填体,密度甚大,全干缩率大,弦、径向干缩比差异也大,水分传导性差,干燥速度缓慢,干燥时易发生表裂、内裂、翘弯和扭曲等现象,属难干材。通过试验,提出了克服干燥缺陷的一些措施,为制定四种木材干燥工艺提供了依据。     

基于FCM的木材X射线图像缺陷分割

在用X射线检测木材缺陷的过程中,很重要的一步就是对图像中的缺陷的提取和分割,其处理的效果直接影响到后面的缺陷识别的正确性。主要是针对木材X射线图像对比度差的特点和固有的模糊性,利用模糊C均值聚类的方法对木材X射线图像进行分割处理,为后续的提取出木材缺陷工作打下基础。     

基于计算机的木材特征提取和分类识别技术研究综述

木材由于内部结构和组成成分的差异，使不同种类木材表现出完全不同的理化性质，并决定其不同的用途和商业价格，因此针对木材的分类识别研究具有重要的应用价值。木材分类识别通常经过木材特征提取和基于特征的分类识别这2个步骤。目前木材特征提取主要利用计算机视觉、光谱分析等技术。木材分类识别是基于木材特征的数字化，这一部分可利用计算机算法实现自动识别，较以往人工识别可大幅提高准确度。文中通过分析近20年来木材特征提取和分类识别的相关文献，介绍各种基于计算机的木材特征提取与分类识别技术的特点及适用范围，并结合计算机技术的发展方向探讨木材特征提取与分类识别技术的发展趋势，以期为构建更准确的木材分类识别技术提供参考。     

木质材料激光表面处理研究进展

激光以其能量密度高、运行轨迹自如、方向性好等优点,被广泛用于木质材料切削加工以及表面处理等领域。其中,木质材料激光表面处理,即利用激光热/光电子效应促进材料发生物理、化学变化,以实现改性的目的。笔者对木质材料激光表面处理原理、激光类型与用途以及激光表面处理技术特点与应用领域等内容的研究现状进行了综述与分析。现有研究表明:材料对于不同波长激光的吸收系数差异明显,基于此特性可实现木质材料表层激光选择性消融,以满足不同应用场合的加工需求;其次,激光功率、激光扫描路径密度以及进给速度等激光工艺参数,决定了表面处理过程中能量密度及分布,进而影响木质材料表面改性强度与深度,通过对激光参数的优化可实现木质材料表面部分性能(如表面颜色、润湿性、功能性结构等)的精准改性。最后,还探讨了木质材料激光表面处理领域研究现状与趋势,阐明了研究的科学意义和未来研究方向,以期为创新木质材料表面处理技术、实现其提质增效提供理论指导,也为拓宽激光技术在木质材料加工领域的应用范围提供思路。     

基于BP神经网络的木材表面颜色特征分类的研究

本文针对木材表面颜色自动分类难题,在 RGB 颜色空间提取木材图像的颜色矩作为颜色特征参数,利用 BP 神经网络对特征参数进行分类,通过输入层、输出层和隐含层的设计,传递函数的选择,确定最终网络结构。实验结果表明,分类正确率达到98%,验证了本文提取的特征参数的有效性。     

无机质复合速生材研究现状与展望

速生人工林种植面积大,生长周期短,在取代天然林资源方面具有得天独厚的产量优势,但存在结构疏松、密度小、强度低等缺陷,导致其使用范围大受限制。使用无机材料对人工速生材进行改性处理可有效提高速生材性能和用途。文中简要总结无机质复合木材制备方法——扩散法、溶胶—凝胶法和真空浸渍法,以及无机材料改性处理对速生木材在力学性能、尺寸稳定性、阻燃性、表面疏水性和表面耐候性方面的改善,以期为人工林速生材无机材料改性提供技术参考。     

基于数字图像处理技术实现木材横纹压缩过程中构造学形态特征的研究

根据计算机视觉技术,结合横纹压缩木材构造形态特征,提出了适合于木材横纹压缩过程中构造形态特征的图像处理方法。对图像质量的完善较全面地进行了量化,充分表露出计算机视觉技术在实现木材横纹压缩过程中构造学形态特征动态变化分析中的优越性,实现了木材细胞边缘轮廓的跟踪,形态特征的提取,为木材科学的再发展奠定了理论基础。     

外负载TiO2功能化改性木材研究现状

TiO₂作为一种新型无机纳米改性剂,对于改善木材尺寸稳定性差、易燃、易朽、不耐虫蛀、易变色、耐老化性差等缺点,赋予其新的功能性具有良好的效果。文中主要对比了TiO₂负载木材的方法,即表面涂覆、水热法、溶胶—凝胶法和超声波辅助溶胶—凝胶法的机理和优缺点;分析了TiO₂改性对木材内部构造及微观结构的影响,归纳了TiO₂改性对木材吸湿性和力学性能的影响;概述了TiO₂在防水、阻燃、防腐、耐候、防变色、光催化和亲疏水双面功能型木材及磁吸附木材中的应用研究现状;归纳总结了TiO₂在改性木材科学研究中存在的问题,并提出合理建议,以期为实际生产和科学研究提供参考。     

Propagation behavior of acoustic wave in wood

We used acoustic tests on a quarter-sawn poplar timbers to study the effects of wood anisotropy and cavity defects on acoustic wave velocity and travel path, and we investigated acoustic wave propagation behavior in wood. The timber specimens were first tested in unmodified condition and then tested after introduction of cavity defects of varying sizes to quantify the transmitting time of acoustic waves in laboratory conditions. Two-dimensional acoustic wave contour maps on the radial section of specimens were then simulated and analyzed based on the experimental data. We tested the relationship between wood grain and acoustic wave velocity as waves passed in various directions through wood. Wood anisotropy has significant effects on both velocity and travel path of acoustic waves, and the velocity of waves passing longitudinally through timbers exceeded the radial velocity. Moreover, cavity defects altered acoustic wave time contours on radial sections of timbers. Acous-tic wave transits from an excitation point to the region behind a cavity in defective wood more slowly than in intact wood.