原木端面图像检尺直径识别算法的研究

论述应用计算机图像处理和识别技术来研究不规则类圆形目标的识别问题,以成捆原木端面图像为实例,提出一种原木端面检尺直径识别的算法。     

基于YCbCr和Hough变换圆的林区原木运输车辆识别

正确识别林区原木运输车辆能有效防止原木被违法运输的异常行为,提高监控管理森林资源的效力。为了解决由于林区道路场景的复杂性,原木端面颜色受光照、湿度等影响使得原木运输车辆识别率较低的问题,将基于YCbCr颜色空间和Hough变换圆检测相结合来识别林区原木运输车辆。同一捆原木端面颜色差异较小,可使用YCbCr颜色特征空间来分割图像,去除背景干扰;图像被转换到RGB空间以去除原木区域二值图像的背景;利用形态学方法统一去除二值化图像的原木缝隙来确定图像边缘;利用Hough变换圆的点线间的对偶性来检测原木运输车辆,降低了噪声的敏感性。实验结果表明,上述方法对成捆裸露在外的原木运输车辆识别率达到了71%以上,鲁棒性和有效性较好。     

一种基于计算机视觉的原木端面面积检测方法

端面面积是原木材积检测中必须检测的参数,直接关系到材积检测的精度和效率。现有的端面面积检测方法以圆形或者椭圆来替代不规则的原木端面,不能直接检测形状不规则原木端面的面积,造成了原木材积测量值与实际值的差异。本研究提出一种基于计算机视觉的原木端面面积检测方法,实现对不规则原木端面面积的直接测量。研究利用该方法对成堆原木的端面面积进行了检测,并与传统的人工检尺法进行了对比,结果表明本研究提出的原木端面检测新方法的检测精度提高了2.8倍、检测速度提高了10~20倍。     

原木端面面积的数字图像检测方法研究

原木端面面积的自动检测是实现原木材积自动化操作的关键步骤。采用数字图像处理的方法,通过灰度变换、中值滤波、边缘增强、二值化和轮廓跟踪等技术处理,实现对原木端面面积的自动检测。试验证明,采用以上方法和技术处理可获得较准确的原木端面面积的检测结果。     

贮木场原木楞堆图像计数技术

贮木场楞堆的计数往往采用人工的方法,而人工计数又存在很多局限性.为解决这一问题提出了一种基于距离变换的图像识别方法.该方法采用中值滤波法、二值化分割和数学形态处理运算实现棒线材图像的预处理,根据距离变换的思想,采用距离变换图中搜索种子点的方法实现原木的计数.该技术尤其适用于贮木场原木楞堆端面摆放参差不齐、原木端面面积大小不一和原木形状不规则的等情况下计数,真有一定的推广和应用价值.     

基于深度学习与深度信息的原木材积检测方法

针对工业现场整车木材摆放深浅不一且上下分层而导致的整车木材材积自动检测困难的问题,采用ZED2双目相机获取木材端面图像以及深度图像,在改进的基于掩模区域卷积神经网络(Mask region-based convolutional neural network, Mask R-CNN)实例分割模型完成对木材端面分割的基础上,对木材轮廓进行椭圆拟合,根据木材端面中心坐标以及深度图像获取每根木材的深度信息,利用整车木材上下层深度信息相差较大的特点,解决木材上下分层的问题,同时根据双目视觉视差原理,利用相机参数以及轮廓拟合后椭圆短径端点坐标,测量出木材端面检尺径,将检尺径和检尺长代入原木材积计算公式完成原木材积的检测。实验结果表明,使用本方法计算出的整车原木材积同人工检尺数据对比,材积误差率(偶数)为-1.61%,单根木材检尺径平均绝对误差(偶数)为0.33 cm,符合国家原木检尺要求。最终将模型以及材积计算算法部署于Nvidia Jetson Xavier NX平台,并在PyQt5核心库上设计了操作界面,用户仅需拍摄完整木材端面,即可实现原木材积的计算。利用本研究所提出方法可实现木材整车检尺...     

原木形状识别技术的探讨

在总结了国内外原木形状识别技术理论发展状况的同时,对我国原木识别技术的发展趋势做了分析,指出利用先进的原木识别技术实现制材优化,提高原料利用率,降低成本,是制材行业发展的必然。     

基于六点定心理论的原木端面计算机视频定心检测理论应用研究

在旋切智能上木定心机器人实验台的定心作业过程中 ,定心数据采集主要包括原木四个截面、两种方式的定心处理 ,其中原木两端面的定心数据采集采用计算机视频检测方法。本文依据旋切原木六点定心理论 ,建立了原木端面视频数据采集的数学模型 ,满足了两端面的定心精度要求 ,同时分析了该理论的应用领域     

原木形状识别机的设计

从原木外形轮廓与制材生产之间的关联性出发，探讨和研究了原木形状识别机设计的原理及设计的方法，为原木形状识别机的设计和制造提供了理论依据。     

基于图像处理的贮木场楞堆原木计数算法

针对楞堆计数的特定问题,提出了一种新的图像识别方案.该方法用二值化分割、图像链码跟踪和直线填充的方法实现原木和背景的分离,根据距离变换的思想,采用距离变换图中搜索种子点的方法实现原木的计数.该算法尤其适用于原木端面形状不规则、面积大小不一、摆放不整齐的楞堆计数.     

一种结合DWT和DCT的彩色图像水印算法

随着计算机网络和多媒体技术的发展,数字水印技术正逐渐成为多媒体版权保护的一种有效途径.数字水印处理的目的是将秘密、个人信息隐含于数字媒体信号中,以保护其知识产权或证明其真实可靠性.提出了一种结合离散小波变换和离散余弦变换的彩色图像数字水印算法.首先,原始彩色图像的绿色分量进行分块DCT变换,选取每块的直流分量组成新的图像,对新图像进行离散Haar小波变换.然后二值水印图像经过置乱加密后嵌入到低频子带.该算法具有较好的透明性和鲁棒性.实验结果表明,对诸如叠加噪声、JPEG压缩、剪切、平滑滤波、图像增强和马赛克效果等普通的信号处理均具有较强的顽健性.     

Numerical and experimental study of moisture-induced stress and strain field developments in timber logs

When solid wood dries from a green condition to a moisture content used for further processing, moisture-induced fracture and stresses can occur. The drying stresses arise because of internal deformation constraints that are strongly affected by the cross-sectional moisture gradient differential shrinkage and the inhomogeneity of the material. To obtain a better understanding of how stresses develop during climatic variations, the field histories of stresses (and strains) in cross sections in their entirety need to be studied. The present paper reports on experiments and numerical simulations concerned with analysing the development of strains and stresses during the drying of 15-mm-thick discs of Norway spruce timber log. The samples were dried at 23 °C and relative humidity of 64 % from a green condition to equilibrium moisture content. The moisture gradient in the longitudinal direction was minimised by use of thin discs simplifying the moisture history of the samples studied. The strain field history was measured throughout the drying process by use of a digital image correlation system. Numerical simulations of the samples agreed rather well with the experimental strain results obtained. The stress results also indicated where in the cross section and when fractures could be expected to occur during drying. More optimal drying schemes showed markedly reduced stress generation.     

基于包络理论的原木截面检测及再现方法研究

介绍了一种应用激光检测技术将原木截面形状再现并计算出原木截面面积的新方法,即应用基于包络理论的原理建立测量系统模型,应用无线数据采集系统采集检测到的数据,然后利用采集到的数据在计算机中进行原木形状的再现,并进行包络线上点坐标的确定以及包络面积的计算,最后对可能出现的结果进行了分析。     

长条薄片刨花三维重建方法研究

提出了根据长条薄片刨花切屑数字图像对其进行三维重建的方法.该方法采用阈值法,以色调为阈值对原始图像中的刨花切屑区域进行分割,利用临域法将噪声区域去除,并针对刨花切屑轮廓多为凹多边形的特点,提出了直接生成刨花切屑表面轮廓三角形网格的算法.试验表明,该方法具有算法简单、重建速度快和效果好等特点.     

0.5M分辨率航摄数字影像数据在森林资源规划设计调查中的应用

以0.5 M分辨率航摄数字影像数据为基础,结合以往调查设计及经营资料,对森林资源规划设计调查小班的部分因子进行判读,探讨高分辨率航摄数字影像数据在森林资源规划设计调查中的应用,结果表明:利用0.5M分辨率航摄数字影像数据进行小班因子判读,可以作为小班因子调查的辅助手段,并可显著提高小班调查质量及效率。     

Intrafiber distribution of aluminum components in alum-treated handsheets

The distribution of aluminum components in cross sections of pulp fibers for alum-treated handsheets was successfully measured by scanning electron microscopy (SEM) and energy dispersive X-ray (EDX) analysis. In this case, gas-phase osmium coating of paper samples is necessary for the samples to have improved stability to long-term irradiation by electron beams at high magnification during the SEM-EDX measurements. The EDX line-analysis method was superior to the elemental mapping image method in the quantification of aluminum components. According to the SEM-EDX measurement for handsheets prepared from a pulp suspension by adding 2% (on dry weight of pulp) aluminum sulfate, the aluminum components were distributed almost homogeneously in the cross sections of pulp fibers. Aluminum species with smaller sizes than those of A1(OH)₃ flocs could probably penetrate the pulp fibers. Thus, aluminum components present only at the pulp fiber surfaces may contribute to retention of colloidal substances in pulp suspensions. From this aspect, a large amount of aluminum components present in the inside pulp fibers may be wasted.This report was presented at the 49th annual meeting of the Japan Wood Research Society at Tokyo, April 1997     

A rapid staining‐assisted wood sampling method for PCR‐based detection of pine wood nematode Bursaphelenchus xylophilus in Pinus massoniana wood tissue

For reasons of unequal distribution of more than one nematode species in wood, and limited availability of wood samples required for the PCR‐based method for detecting pinewood nematodes in wood tissue of Pinus massoniana, a rapid staining‐assisted wood sampling method aiding PCR‐based detection of the pine wood nematode Bursaphelenchus xylophilus (Bx) in small wood samples of P. massoniana was developed in this study. This comprised a series of new techniques: sampling, mass estimations of nematodes using staining techniques, and lowest limit Bx nematode mass determination for PCR detection. The procedure was undertaken on three adjoining 5‐mg wood cross‐sections, of 0.5 × 0.5 × 0.015 cm dimension, that were cut from a wood sample of 0.5 × 0.5 × 0.5 cm initially, then the larger wood sample was stained by acid fuchsin, from which two 5‐mg wood cross‐sections (that adjoined the three 5‐mg wood cross‐sections, mentioned above) were cut. Nematode‐staining‐spots (NSSs) in each of the two stained sections were counted under a microscope at 100× magnification. If there were eight or more NSSs present, the adjoining three sections were used for PCR assays. The B. xylophilus– specific amplicon of 403 bp ( DQ855275 ) was generated by PCR assay from 100.00% of 5‐mg wood cross‐sections that contained more than eight Bx NSSs by the PCR assay. The entire sampling procedure took only 10 min indicating that it is suitable for the fast estimation of nematode numbers in the wood of P. massonina as the prelimary sample selections for other more expensive Bx‐detection methods such as PCR assay.     

Plywood frame corner joints with glued-in hardwood dowels

This paper presents research on plywood frame corners jointed to glulam beams and columns by means of glued-in hardwood dowels. The frame corner was made of a solid block of ordinary plywood of the same width as the glulam beams and with plies parallel to the plane of the frame to avoid splitting due to stress perpendicular to the grain. Hardwood dowels with a diameter of 12mm and a maximum glued-in length of 120mm were glued into drilled holes in the plywood corner and glulam beam ends parallel to the grain direction of the beams to form a momentresisting joint. Static bending tests were conducted of frame corners with 100 X 200mm² and 120 X 420mm² beam cross sections. Bending capacities of the joints corresponding to a modulus of rupture of the jointed glulam beams of about 30MPa were obtained for both closing and opening moments for the small cross sections, and about 22MPa was obtained for the large cross sections. Simple design models for calculation of joint strength and rotational stiffness are also presented.     

数字图像技术在木材科学中的应用

随着现代计算机技术的快速发展,计算机数字图像处理技术已成为当前追逐的热点,并广泛应用在木材科学领域中。数字图像技术所具备的现性好、精度高、灵活性强等优势,为木材科学的发展提供了新手段。文中基于数字图像处理技术从木材宏观与微观方面对木材缺陷、木材纹理、细胞特征等进行综述,同时分析了木材科学研究中主要用到的数字图像处理技术,并对数字图像处理技术在木材科学中的研究方向及发展趋势进行展望。     

基于细胞数字特征的板材材种识别技术

木材识别是合理利用木材最基本的前提。为了解决人工鉴定失误率高的问题,本文提出基于细胞数学描述理论的板材识别方法,采用木材端面细胞的显微数字图像,利用板材材种的数学仿真技术、图像处理技术、数据库管理和系统辨识方法,提取细胞轮廓形态和尺寸等参数建立基准细胞和材种识别数字化参数库。减少传统的依赖图像像素特征识别比较方法的不确定性,创新材种识别的数字化理论体系,推动板材材种识别的实用化。