基于邻接自适应谱聚类的木材表面缺陷分割算法

针对人工分割木材表面缺陷的烦琐性和阈值分割算法对缺陷像素信息衡量的不稳定性,提出了一种基于邻接自适应谱聚类的木材表面缺陷分割算法。算法以简单线性迭代超像素(simple linear iterative cluster, SLIC)为基础,对缺陷图像进行预处理,融合木材缺陷的纹理特性和超像素块间的距离尺度,并采用邻接自适应谱聚类进行分割;缺陷分割初步完成后,通过变异系数衡量缺陷块中像素信息的离散程度进行再次分割,克服初次分割结果的过分割问题;考虑木材表面缺陷形态学上的封闭性,将2次分割图像进行合并,继而用邻接扫描法对次分割图形进行填充,最终对木材表面缺陷进行分割界定。考虑木材表面缺陷种类的多样性,选取了虫眼、死节、活节等缺陷图像进行分割对比试验,相较于OTSU阈值分割算法,本研究算法在单个和多个木材表面缺陷分割方面,类别平均像素准确度(mean pixel accuaracy, MPA)分别提升4.69%,14.23%,平均交并比(mean intersection over union, mIoU)分别提升33.27%,33.43%。本研究算法能够更加准确地将木材表面缺陷从复杂背景中...     

人造板表面缺陷检测图像自适应快速阈值分割算法

【目的】提出一种自适应快速阈值图像分割算法,为人造板表面缺陷在线检测提供支持。【方法】首先将整幅图像划分成若干子区域,通过计算子区域的方差对缺陷进行定位,提取出缺陷所在区域,只对缺陷区域进行图像分割,解决小面积目标难以准确分割的问题。然后对缺陷区域的一维灰度直方图进行处理,直方图平滑后去除掉不显著波峰,根据处理后保留的主要波峰数量和位置自适应地确定分割阈值个数以及每个阈值的分割区间,实现当图像中出现多种类型缺陷时算法自动确定分割阈值个数。最后,通过分析Otsu算法,将阈值穷举搜索改进为条件搜索并限定搜索方向,在每个分割区间内使用改进的Otsu算法对阈值进行搜索,提高搜索速度。【结果】对板面存在油污、大刨花、胶斑、杂物、松软5种类型缺陷的人造板表面图像进行分割,在板面缺陷数量、类型不固定的情况下,算法可以自适应地确定分割阈值个数,在15 ms内将各种类型缺陷从人造板表面图像中分割出来,平均分割准确率达97%。【结论】自适应快速阈值分割算法能够快速、准确将缺陷从人造板表面图像中分离出来,在执行速度和分割效果上均满足在线缺陷检测系统的要求,可为人造板表面缺陷在线检测提供新思路。     

基于图论的图像分割算法分析研究

图像分割是把图像分成各具特性的区域并提取出感兴趣目标的技术和过程,图像分割新算法一直在更新,但都有局限性。本文介绍几种基于图论的图像分割算法,并对各自算法特点进行分析,通过将Normalized Cut归一化割集准则与图像的阈值分割联合起来区分目标和背景,提出了改进权值公式的算法。通过实验比较分析,改进的归一化图像分割算法有效消除了噪声,取得良好的实验效果,更接近人眼视觉的分割效果,并减少了算法的运算量。     

基于图像检索的林火检测技术研究

基于计算视觉的常规林火图像识别算法过多依赖于庞大训练集,算法复杂无法满足林火监测实时性要求,提出一种基于图像检索的林火识别算法,搭建了VS2013和OpenCV程序验证平台。采用拉伸直方图和直方图均衡化方法提高图像对比度,通过减色降低程序计算复杂度,设定林火参考图像和固定阈值,采用直方图对比算法检索林火图像。程序验证结果表明,本文设计的程序能快速识别林火图像,具有实用价值,可为相关研究提供参考。     

机载激光雷达人工林单木分割方法比较和精度分析

【目的】研究分水岭算法、四次多项式拟合法和基于点云的距离判别聚类法对人工林单木分割的适用性,分析3种方法对人工林单木分割的精度,探索进行单木分割时3种方法关键参数的最优选择。【方法】结合地面实测数据和目视解译方法,计算单木探测率、准确率和F得分,比较分水岭算法、四次多项式拟合法和基于点云的距离判别聚类法的单木分割精度,并通过改变栅格化冠层高度模型(CHM)的分辨率及调整基于点云的距离判别聚类法的距离阈值,分别对3种方法进行单木提取效果的敏感性分析。【结果】1)分水岭算法、四次多项式拟合法和基于点云的距离判别聚类法对人工林单木总体分割精度较高(F=0.76～0.83); 2)对于"复杂林型"样地,基于点云的距离判别聚类法的分割精度最高(F=0.78),优于分水岭算法(F=0.74)和四次多项式拟合法(F=0.53);对于"中等复杂林型"样地,基于点云的距离判别聚类法的分割精度最高(F=0.89),优于分水岭算法(F=0.84)和四次多项式拟合法(F=0.75);对于"简单林型"样地,基于点云的距离判别聚类法(F=0.89)、分水岭算法(F=0.89)和四次多项式拟合法(F=0.93)的分割精度都较高; 3)敏感性分析结果表明,当CHM分辨率为0.5 m×0.5 m时,分水岭算法和四次多项式拟合法的分割精度最高;当基于点云的距离判别聚类法的距离阈值近似样地平均冠幅半径时,其分割精度最高。【结论】对多种类型样地进行单木分割,体现了分水岭算法、四次多项式拟合法和基于点云的距离判别聚类法对人工林单木分割的适用性;结合多种类型样地充分评估并比较了3种方法对人工林单木分割的精度;通过对3种方法进行敏感性分析,阐述了进行单木分割时关键参数的最优选择。     

林木幼苗根系CT序列图像分割

图像分割技术一直都是图像处理中的难题,原位根系CT图像的分割结果影响后续的三维重建工作,选取合适的分割算法十分关键.通过对根系CT图像分割算法以及根系CT图像特点的的研究,选择阈值分割、区域生长和FCM聚类的方法对序列图像进行分割,并将区域生长方法进行优化,减少分割时计算的步骤,同时改进f cm聚类方法,将空间信息融入到FCM目标函数中.结果表明,阈值法仅对序列初始的简单图像处理效果理想,改进的区域生长法对分割目标连续的图像分割效果好,分割速度达到0.6s.而改进的FCM虽然需要花费23 s的时间,但对不连续的分割目标分割效果明显.三种方法均体现出分割的准确度,后两者能有效提升分割的效率.适当选取以上三种分割方法,即可快速准确的完成林木幼苗CT序列图像的分割.     

木材缺陷的阈值分割算法研究

阈值分割算法包括迭代阈值分割、最大方差阈值分割、最大熵阈值分割等,使用3种阈值分割算法进行木材缺陷图像分割,多次实验证明,最大方差阈值分割算法的分割效果最好.在此基础上,对最大方差阈值分割算法进行增强,调整灰度级,增大目标和背景区域的对比度.实验表明,该方法可以明显改善最大方差阈值分割的分割效果.     

基于Gabor变换的纹理图像分割算法及应用

基于Gabor小波的多尺度、多方向性,结合模极值特征提取方法和聚类算法提出一种新的纹理图像分割算法,并将算法应用到竹材横端面的识别中。二维Gabor小波可以在空域、频域和方向上获得最佳的分辨率,利用其对纹理图像进行分解,提取图像模极值纹理特征,用聚类算法对纹理图像进行分割。实验结果表明,提出的方法对合成纹理图像有理想的分割效果,并能够很好地对自然纹理图像进行分割。算法在竹材横端面的分割识别中,取得了很好的效果。     

基于FCM和分水岭算法的无人机影像中林分因子提取

【目的】研究高精度小型无人机获取林分调查因子方法,将林分调查因子在低空无人机影像上识别并提取出来,获取树高、冠径等测树因子,建立林分因子测量方法,实现经济、高效、快捷、精准的森林资源调查和监测,及时掌握森林资源及相关林分因子的时空变化特征。【方法】以东北林业大学城市林业示范基地樟子松人工林为研究对象,以多旋翼无人机影像为数据源,基于FCM聚类算法和分水岭分割算法以及形态学运算、阈值分割、图像平滑、灰度化、二值化等一系列数字图像处理技术,提取樟子松人工林林分因子。FCM聚类算法和阈值分割法用于提取树梢标记图像,分水岭分割算法对树梢标记图像进行迭代处理从而获得单木树冠分割图像,根据单木树冠分割结果提取单木特征进而计算各林分因子值。【结果】在林地提取中,根据影像的颜色特征绿度分割成功地将林地部分与非林地部分分离开来,确定单木树冠分割范围。在单木树冠分割中,阈值分割法和FCM聚类算法均可有效将树梢标记从林地图像中提取出来;将基于标记的分水岭分割算法用于单木树冠分割取得较好效果,大多数单木树冠被单独分割出来,但某些区域仍然存在一定的欠分割或过分割问题。在林分因子提取中,提取的林分因子包括林分郁闭度、林地面积、立木株数和平均冠幅,其中林分郁闭度的测量精度为96.67%,林地面积的测量精度为81.23%,立木株数和平均冠幅的测量精度与单木树冠分割中的树梢提取方法(阈值分割法和FCM聚类算法)及分水岭分割中的2个参数(形态学腐蚀的结构元素大小和中值滤波的窗口大小)有关。针对2种树梢提取方法,分别进行参数组合试验,结果显示2种树梢提取方法使用适当参数组合所得各林分因子测量精度均在80%以上,平均测量精度均在90%以上,其中阈值分割法的最高平均测量精度为94.49%,FCM聚类算法的最高平均测量精度为93.17%。【结论】利用无人机拍摄的人工林影像进行森林资源调查,将先进的计算机科学技术和无人机技术应用到林业领域中,可有效提高森林资源调查的效率和精度。本研究提出的林分因子提取方法适用于高郁闭度林分,测量精度满足实际需求。     

基于模糊聚类分析的木材缺陷CT图像分割

为了提高木材的利用率,在木材加工之前对木材缺陷CT图像进行分割,将节子和空洞等缺陷分割出来,通过观察缺陷的位置便于工人师傅下锯。利用计算机断层扫描（CT）技术获取木材缺陷图像,将数字图像处理技术与模糊聚类算法相结合,在标准的模糊C均值算法的基础上改进,采用半模糊聚类的分析方法对木材缺陷图像进行分割检测。实验结果表明：基于半模糊聚类的图像检测方法在木材图像检测上取得了较好的效果,缺陷边缘处很平滑,细节保留完整,更多的保留了边缘上的信息。从而证明了半模糊聚类分析法在木材缺陷CT图像处理方面具有可行性。     

Estimation of area burned by forest fires in Mediterranean countries: A remote sensing data mining perspective

Forest fires throughout the world result in tree mortality that can cause substantial timber and carbon losses. There is a critical need to map the areas burned by such fires to guide forest management decisions. Moderate Resolution Imaging Spectroradiometer (MODIS) satellite imagery provides inexpensive and frequent coverage over large areas, facilitating forest health monitoring. In this study a MODIS post-fire image at a spatial resolution of 250 m serves as the starting point of an image mining based method. It involves three algorithms: modeling as a sum of Gaussian functions, kernel based smoothing, and adaptive thresholding. Adaptive thresholding serves as the reference to be compared to the image mining based method. Three spectral indices specifically designed for burned area identification have been used: the Burned Area Index (BAI), the Burned Area Index adapted to MODIS bands (BAIM), and the Normalized Burn Ratio (NBR). The κ statistic is applied to quantify the accuracy of the burned areas estimations by relating the estimated area with burned area perimeters measured on the ground by Global Positioning System (GPS). In addition, the κ statistic allows us to identify both the optimal spectral index and the optimal algorithms’ parameters. In this work, an accurate estimation (κ > 0.8) of areas burned by forest fires in Mediterranean countries is achieved, in particular if the BAIM index is used. The accuracy of these estimates is compared with the accuracy obtained by using the reference method by a McNemar’s test. Results show that our image mining based method allows a higher accuracy (the average increase of κ equals to 16%) than the reference method. We conclude that this method adequately maps burned areas, and that it may help management agencies to better understand of landscape-scale burn patterns.     

基于灰度-梯度共生矩阵的木材表面缺陷分割方法

根据木材表面缺陷图像的特点，提出了基于灰度-梯度共生矩阵模型和最大熵原理的木材缺陷图像自动阈值化技术。并且利用形态学运算对分割后的二值图像进行后期处理。经实验验证，该方法提取的木材表面缺陷图像效果良好：     

Using mixed integer programming and airborne laser scanning to generate forest management unitsOA

Airborne laser scanning(ALS) has been widely applied to estimate tree and forest attributes, but it can also drive the segmentation of forest areas. Clustering algorithms are the dominant technique in segmentation but spatial optimization using exact methods remains untested. This study presents a novel approach to segmentation based on mixed integer programming to create forest management units(FMUs). This investigation focuses on using raster information derived from ALS surveys. Two mainstrea...     

Characterization of larval instars of artemisia weevil,Adosomus sp.

The artemisia weevil,Adosomus sp.,has become an important wormwood(Artemisia ordosoca and A.sphaerocephala) pest in recent years in Ningxia Province,China.The number of larval instars should be determined before further studies on its life history,biology and integrate management of pest(IMP).In this study,we measured the length and width of head capsule,based on frequency histograms and Dyar's law,from which we concluded that the larvae of Adosomus sp.have four instars.Of the two vari-ables measured,the wi...     

Clear-cut Detection in Boreal Forest Aided by Remote Sensing

The study compares the applicability of different remote sensing data and digital change detection methods in detecting clear-cut areas in boreal forest. The methods selected for comparisons are simple and straightforward and thus applicable in practical forestry. The data tested were from Landsat satellite imagery and high-altitude panchromatic aerial orthophotographs. The change detection was based on image differencing. Three different approaches were tested: (1) pixel-by-pixel differencing and segmentation; (2) pixel block-level differencing and thresholding; and (3) presegmentation and unsupervised classification. The study shows that the methods and data sources used are accurate enough for operational detection of clear-cut areas. The study suggests that predelineated segments or pixel blocks should be used for image differencing to decrease the number of misinterpreted small areas. For the same reason the use of a digital forest mask is crucial in operational applications.     

Computer Simulation of Fiber Length and Width Distribution for Two Poplar Woods

1 INTRODUCTIONPoplar is the major afforestation species in China, andhas been used for pulping since 1960s (Xu Weiying,1988). Fiber morphology makes remarkable effectson quality of paper, and attract great attention (KurtH M, Lynch D L, 2001).The quantification, via modeling, of the variation inwood properties is usually accomplished usingstatistical techniques. Fundamental of theseapplications is a variety of probability distributions(Melo, et al, 2000). Due to considerable degree of…     

Method of check analysis of microwave-modified wood

A photo-based method for statistical image analysis of microwave (MW)-modified timber was developed and applied to test sets of Radiata pine and eucalypt hardwoods (Messmate and Mountain Ash). The method is based on filling the checks in timber with stain solution and analysing the wood surface. Bethel impregnation process (in a pressurised treatment vessel) was used to produce uniform surface staining of the test samples. Image processing was automated and the number of manual operations (the human decision-making process) was minimised. A computer program that automates thresholding and a program that repairs the threshold image were written. The software allowed larger images to be analysed and reduced image processing time. The described method produced reliable check measurements and statistics for softwoods. Though suitable for estimating the quality of individual MW-modified hardwood specimens, the method was found to be inadequate for statistical analysis of eucalypt hardwoods because of collateral staining of vessels accompanied with a high level of stain penetration into the wood tissues.     

配送中心线路优化方法的探讨

基于目前配送中心线路优化的各种方法,根据实际情况对上述算法进行分析和比较,并针对模拟退火法、遗传算法在配送线路优化问题中的具体应用,提出不同情形下的求解策略。     

不同森林经营措施对木材产量和碳储量的影响

在应对全球气候变化中, 林业具有特殊地位和重要功能, 以社会、经济和环境可持续发展为目标的森林可持续经营是全球趋势。文中阐述了森林在减缓和适应气候变化中的独特作用; 重点分析了不同营林措施对木材产量和碳储量的影响, 并结合可持续林业的发展趋势评价了不同经营模式框架下木材产量与碳储量之间的平衡; 针对全球气候变化对森林生态系统带来的挑战, 探讨了今后森林经营措施的研究重点和难点, 旨在为制定合理的经营方案提供理论依据。     

Comparison of spatially explicit forest landscape fire disturbance models

Comparisons of model behaviors are an efficient way of understanding the differences amongst various models and thus providing guidance to model users for selecting suitable models for their own purposes. This study focuses on a comparison of the most commonly used fire spread algorithms used in scenario fire regime models. This paper provides an overview of fire regime modeling and describes a simulation model, Ecological Disturbance Model, as a simulation shell for such a comparison using the Fort A La Corne forest area in central Saskatchewan, Canada, as the study area. Simulation results suggested that for a fire scenario modeling approach, various fire spread algorithms such as DISPATCH, percolation, and cellular automata (CA) may not result in significant differences between user-defined and simulated fire frequencies; however, significant differences in simulated forest dynamics could result when using different fire spread algorithms. The simulation results from DISPATCH and CA are more similar than those from the percolation algorithm; however, the latter appeared to be a better representative of observed fire spread processes due to its underlying assumption of fire spread mechanisms. Simulation results also suggested that the fire spread algorithms that replicate four or eight direction fire spread in percolation and CA will not make a significant difference in simulated fire regimes or forest dynamics. It is thus recommend that using simulation shells as a tool to take alternative assumptions or models into account to narrow down the uncertainty parameters and avoid the paradoxes in the modeling of natural resource management.