节点文献

北京市城六区三维绿量估算与分析研究

Estimation and Analysis of Tridimensional Green Biomass of Central Six Districts of Beijing

【作者】 唐雪海

【导师】 冯仲科; 黄庆丰;

【作者基本信息】 北京林业大学 , 林业装备工程, 2011, 博士

【摘要】 北京市城六区三维绿量估算与分析研究是区别以往的根据树冠形状选择合适的体积公式来计算树冠的体积,利用地面三维激光扫描系统对单株立木进行扫描,根据树冠的外部点云数据计算体积,得到单株的三维绿量。结合RS和GIS技术,借助数理统计知识,建立三维绿量反演模型,得出北京市城六区的三维绿量并进行相关分析,提出意见和建议,为城区绿地规划营造出能发挥最佳生态效益和其他综合效益的绿色空间提供科学依据。本文基于地面三维激光扫描系统对北京市城六区的针叶树和阔叶树进行单株扫描,获得单株的三维绿量,并采用SPOT5遥感影像作为数据源,借助SPSS多元统计软件和遥感及地理信息系统建模及空间分析工具,建立三维绿量模型,对北京市城六区的森林植被三维绿量进行了估算,并进行了相关分析,为快速准确地估算北京市城六区森林植被三维绿量总量提供了一种有效的途径和方法。得出如下结论:(1)通过系统抽样和典型抽样,在校园、公园、路旁、住宅小区、停车场等绿化场合运用地面三维激光扫描仪FARO LS880对北京市常见的三十余种绿化树种如侧柏、雪松、白皮松、油松、圆柏、银杏、杨树、梧桐、柳树、国槐、臭椿、栾树、白蜡、元宝枫、栓皮栎等进行单株扫描,每个树种扫描三十余株,共计扫描1000余株。通过FARO SCENE点云处理软件分别提取单株立木的树冠点云,并在逆向工程软件Geomagic中分别量测每株扫描数据的树高、胸径、冠幅、枝高等测树因子;通过程序编写读取单株树冠各个点云的相对三维坐标,分别计算垂直于树干方向多个面上由最外围点构成的不规则面的面积,然后在沿树干方向等距切割计算近似圆台的体积,最后累加得到树冠体积,即获得单株立木冠上部分的三维绿量;(2)选择空间分辨率为2.5m的全色波段和空空间分辨率为10m的多光谱波段的SPOT5遥感影像数据,结合Worldview和Google earth影像及地形图等辅助数据,采用监督分类和目视解译相结合的方法对北京市城六区的植被信息进行分类与提取,本文共分了针叶林、阔叶林、草地、农田、水体、建筑用地和其他用地六个地类,总体分类精度达到83%,通过实地调查点的验证,相符率超过93%;通过多个绿化树种的地面扫描数据,结合遥感和GIS信息,应用多元统计分析方法进行自变量的筛选,分别建立针叶树和阔叶树的三维绿量模型,其中针叶树的三维绿量模型是:V=1.149-0.096B1-0.182+0.199SWIR,阔叶树的三维绿量模型是:V=-40.290-0.236B1-0.188B2+0.487SWIR,模型的相关系数为0.946和0.893,显著性显示为极显著,通过回归标准残差和累计概率分布图可以看出模型回归较好,利用检验样本数据与回归估算数据在0.05和0.01置信水平下进行精度评价,结果得知针叶树的精度为96.85%、95.84%,阔叶树的精度为89.29%、85.84%,可以得出利用以上方法建立的三维绿量估算模型是可以使用的。(3)由三维绿量模型估算出北京市城六区的总绿量,通过分析得到北京市城六区的总绿量为399129501.69 m3,其中海淀区最大,为141087038.92 m3,西城区最小,为9943012.58 m3;单位面积绿量最大的是石景山区,每平方米0.74 m3,最小的是朝阳区,每平方米仅0.14 m3,北京市城六区单位面积绿量为每平方米0.29 m3;人均绿量最多的是石景山区,每人105.79 m3,最少的是西城区,每人仅为6.06 m3;在生态效益方面,北京市城六区年产氧量为804981.18t,二氧化碳的年吸收量为1109914.57t,二氧化硫的年吸收量为1209.36t,年滞尘量为439042.45t,其中北京市城六区二氧化硫的年吸收量仅为排放量的1.02%,年滞尘量为降尘量的7.11倍,整体看来北京市城六区绿化植物的生态效益能对环境改善起到一定的作用。通过分析,从增加乔木数量、立体绿化、完善绿地规划体系和制定三维绿量衡量标准三个方面提出建议。本文的创新点主要体现在以下几个方面:(1)近些年,众多学者对三维绿量进行了大量的研究,但大多都是由实测树木的胸径、冠幅、树高等测树因子通过树冠的冠形形状选择合适的体积公式进行计算,这种计算过于粗放,本文通过地面三维激光扫描仪对单株树木进行不同方向的扫描来获得整株树木的完整点云数据,在后处理软件中生成树冠的三维点云模型,分别从沿树干方向和冠幅平面进行等距离分割,在冠幅平面方向又沿水平方向等距离分割,生成若干个梯形和首尾两个三角形,然后计算由点云坐标围成的不规则面的面积,最后按照极小等距形成的若干个圆台和两个圆锥来计算整个树冠的体积,整个过程由程序实现。通过这种方法最后得出的结果去除了模拟规则形状而参与计算的多余空间的体积,结果更加精确,其研究方法具有创新性。(2)本文以全站仪获得的单株立木的坐标为原点,树冠直径的二分之一为半径建立树冠冠径的缓冲,然后通过提取不同缓冲范围在影像各波段中的灰度值建立实测三维绿量与波段值及其组合值之间的模型。这种方法对于基于像元灰度值与实测三维绿量建立模型是一种尝试和探索,对于研究灰度值的提取及三维绿量反演模型建立具有理论意义。(3)对由地面三维激光扫描仪获得的单株树木的点云数据通过Geomagic软件量测冠径、树高、枝高、胸径等测树因子,再对经典的Logistic冠径-冠高方程进行修正,由模型检验和精度评价来反映模型本身的回归显著性,再用实测值与估算值进行精度评价,进而应用到其他扫描树种中,得到多种北京市城六区常见的绿化树种的茎-高模型,此法对于研究城区森林植被的茎高关系具有实践价值和推广意义。

【Abstract】 ABSTRACTIn this paper, tridimensional green biomass of urban six districts(Xicheng, Dongcheng, Haidian, Chaoyang, Shijingshan and Fengtai) of Beijing has been estimated and analyzed. Different from the conventional method that proper volume formula is selected according to the crown shape, single tree scanning by terrestrial three dimensional laser scanning system has been used, and peripheral point cloud data of crown was used to calculate the volume and obtained tridimensional green biomass of single tree.Combined with the RS and GIS technology, the three dimensional green biomass inversion model has been established with the help of mathematical statistics knowledge. The three dimensional green biomass of urban six districts was got and analyzed. In addition, suggestions have been proposed to provide the scientific basis for urban green land planning.Based on the terrestrial laser scanning system, conifers and broad-leaved trees of urban six districts of Beijing have been scanned plant by plant, to acquire tridimentional green biomass per plant. SPOT5 remote sensing was used as the data source, the three-dimensional green biomass model has been established with the help of multivariate statistical software SPSS and remote sensing and geographic information systems model and spatial analysis tool. The three-dimensional green biomass of forest in urban six districts of Beijing was estimated and the correlation analysis has been done, in order to quickly and accurately estimate three dimensional green biomass of urban six districts of Beijing, providing an effective means and methods. The following conclusions have been got:(1) Through systematic sampling and typical sampling, selecting several greening places such as campus,parks,sidewalks,residential districts,parking lots etc, in these places, terrestrial laser scanner FARO LS880 was used to scan over 30 tree species plant by plant, like arborvitae, cedar, pine, cypress, ginkgo, poplar, sycamore, willow tree, Sophora japonica, Ailanthus altissima, Koelreuteria, ash, maple, cork oak and other nearly ten common tree species in Beijing, and a total of more than 1,000 trees have been scanned. By using FARO SCENE point cloud processing software, single plants were extracted of plant crown point cloud, and in the reverse engineering software Geomagic, the factors like height, diameter, crown width, branch were measured plant by plant; By reading through the programming, relative three-dimensional coordinates of the point cloud concerning the crown of each plant were measured and then the outermost point of the surface formed by the irregular surface area of direction perpendicular to the trunk was calculated, and then along the direction of trunk, the infinitesimal method was used to cut a round table volume, and finally they were added together to acquire the crown volume, that is the tridimentional green biomass of crown upper;(2)The image of SPOT5 with the spatial resolution of 2.5m panchromatic band and an empty spatial resolution of 10m multispectral remote sensing data has been selected, combined with Worldview and Google earth and auxiliary data like topographic maps etc, and then combination of supervised classification and visual interpretation methods has been used to do classification and extraction of the vegetation in the urban six districts of Beijing. In this paper,6 land types have been classified:coniferous forest, broadleaf forest, grassland, farmland, water, construction sites and other lands, and the overall classification accuracy reached 83%. Verified by field survey points, conformity was over 93%; According to the scanning data of tree species, remote sensing and GIS information, multivariate statistical analysis methods was used for variable selection, and then three dimensional green biomass models of coniferous and broadleaf trees have been established. The coniferous trees three-dimensional green biomass model is:V=1.149-0.096B1-0.1B2+0.199SWIR, and broadleaf trees three-dimensional green biomass model is: V=-40.290-0.236B1-0.188B2+0.487SWIR. with the correlation coefficient of models were 0.946 and 0.893 respectively, and the significance was shown as highly significant. Through regression standard residual and cumulative probability distribution, the regression of model was good, and the test sample data was used for accuracy estimation under the regression estimates in 0.05 and 0.01 confidence level. The results showed that the accuracy coniferous trees were 96.85% and 95.84%, and the accuracy of broad-leaved trees were 89.29% and 85.84%. Therefore, it can be drawn that the tridimensional green biomass established by the above method can be used.(3) The three-dimensional green biomass model has been used to estimate the green biomass of urban six districts of Beijing. By analysis, the total amount of green biomass concerning urban six districts of Beijing is 399,129,501.69 m3. Haidian District owns the largest amount of 141,087,038.92 cubic meters, Xicheng District owns the least of 9,943,012.58 cubic meters; The largest amount of green biomass per unit area is in Shijingshan District, with 0.74 cubic meters per square meter, and the smallest is in Chaoyang district, with only 0.14 cubic meters per square meter. As for the urban six districts of Beijing, the amount of green biomass per unit area green 0.34 m3/m2; the largest amount of green biomass per capita is in Shijingshan District, with 105.79 cubic meters per capita, while the Western District owns the least green biomass with only 6.06 cubic meters per capita; To ecological benefits, the forestry can product 804981.18t oxygen annual in six districts,the annual absorpted carbon dioxide 1109914.57t, the annual absorpted sulfur diocide 1209.36t,the annual dust reduce quantity 439042.45t, in which the annual absorption tanke up account for 1.02% emission,the annual dust reduce quantity is 7.11 times of dust fall quantity, tthe ecological benefits of green plants in central six districts can play a role in improving the environment in whole. By analysis, suggestions have been proposed from the following three aspects:increase tree number and vertical planting, improve the planning system of green land, set standards of estimating three dimensionnal green biomass as well.Innovation of this paper is shown as follows: (1) In recent years, many scholars have done a lot of research in the three-dimensional green biomass, but for most studies, the volume formula of crown is selected by measuring the factors like tree diameter at breast height, crown width, to estimate the volume. However, this calculation is too extensive, and in this paper three-dimensional laser scanner was used to scan the trees in different directions to the complete point cloud data of the whole trees. And in the post-processing software, three-dimensional point cloud model of the crown has been generated. The model was put into equidistant partition in the direction of the trunk and crown plane, and then, equidistant partition was also done in the direction of the horizontal crown plane, generating several ladder and two drive end to end triangle. In addition, the coordinates of the point cloud bounded by the irregular surface area was calculated. Finally, the crown volume was got based on a number of small round table and two cones, and the whole process was realized by the Program. The results was got by this method, with more accurate results, therefore, the study method is innovative.(2)The paper taking the coordinates of sigle tree obtained by the total station as the origin,one-half the diameter of the crown as radius to establish buffer, and then extracting gray value of the buffers in the image to model.. This method based on pixel gray value and measured tridimentional green biomass is an attempt and exploration, having theoretical meaning in the study of gray value extraction and tridimentional green biomass inversion modeling..(3)The point cloud of single tree got by three-dimensional laser scanner was measured by Geomagic software, to get the measurement factors like crown diameter, tree height, branch height, diameter and others, and then the classical Logistic crown diameter-crown height equation was used for correction. In addition, testing and accuracy evaluation of the model was used to reflect the regression model itself significant, and then measured value and calculated value were used for accuracy assessment. And it was applied to scanning other species to get a Stem-High model with variety of common tree species in urban six districts of Beijing, and this method to study urban vegetation for stem height relations has the practical value and promotion significant.

节点文献中: