基于RGB模式的一种草地盖度定量快速测定方法研究

盖度是研究植物群落结构的一个重要数量指标,传统目视估算法精度低且受人的影响较大。本研究分析了大量荒漠草地实地照片的RGB颜色模式特征,构造了RGB颜色判别决策树区分植被与非植被像元来计算植被的覆盖度,同时在Matlab 7.0平台上将数码照片的导入、地物判别、盖度计算、划分结果显示、对照等功能模块集成,构建用户界面友好、人机交互便捷软件模块,从而能够准确、高效地计算草地盖度,方便了草原实地调查工作。与针刺法相比,该方法计算得到的植被覆盖度最大偏差绝对值不超过5%,精度在95%以上,解决了常规人工方法中估测草地盖度工作量大、不能准确定量测定的问题。     

基于小型无人机与MODIS数据的草地植被覆盖度研究——以甘南州为例

以甘南州为研究区,采用Canon数码相机和小型无人机搭载相机在不同大小样方上获取草地植被覆盖度数码照片,结合2015年5月-10月的Terra/MODIS植被指数产品MOD13Q1,分析了增强型植被指数(EVI)和归一化植被指数(NDVI)与草地植被覆盖度之间的相关性,建立了研究区草地植被覆盖度的回归模型,并对模型进行了精度评价,筛选出甘南州草地植被覆盖度最优遥感反演模型,并对草地生长季时期覆盖度时空上的动态特征进行分析。结果表明,1)利用小型无人机搭载相机获取草地大样方植被数码照片的方法能用于地面草地覆盖度数据的获取;2)与NDVI相比,用EVI估算草地覆盖度更优,因此确定基于EVI构建的对数模型为甘南州草地植被覆盖度最优反演模型,模型精度可达88.00%;3)研究区2015年生长季草地植被覆盖度除了低平地草甸在8月达到最大值外,其它草地类型均在7月达到最大值;4)甘南州以中高植被覆盖度为主,主要分布在玛曲、碌曲、夏河以及合作四县市。整体而言,中西部和西南部区域草地覆盖度高于东部。通过精确草地植被覆盖度模型的建立,不仅有利于及时准确的了解草地植被覆盖度的时空分布状况和季节性动态变化,也有利于维护甘南州草地生态系统的持续稳定发展。     

基于无人机可见光图像的荒漠草地覆盖度估算

及时准确监测草地植被覆盖度,对草地资源的可持续利用及生态系统的恢复与重建具有重要意义。本研究以荒漠草地植被为研究对象,采用监督分类与植被指数直方图相结合的阈值法,分析了6种RGB植被指数对荒漠草地的识别效果。研究结果表明： 归一化绿红差异指数 (normalized green-red difference index, NGRDI)对草地覆盖度的提取精度最高,其平均绝对误差和均方根误差分别为2.56%和3.06%。监督分类与可见光植被指数统计直方图相结合的阈值法对荒漠草地植被覆盖度的提取效果较好,可以用于荒漠草地植被覆盖度的提取。     

基于无人机可见光图像的荒漠草地覆盖度估算

及时准确监测草地植被覆盖度,对草地资源的可持续利用及生态系统的恢复与重建具有重要意义.本研究以荒漠草地植被为研究对象,采用监督分类与植被指数直方图相结合的阈值法,分析了6种RGB植被指数对荒漠草地的识别效果.研究结果表明:归一化绿红差异指数(normalized green-red difference index,NGRDI)对草地覆盖度的提取精度最高,其平均绝对误差和均方根误差分别为2.56％和3.06％.监督分类与可见光植被指数统计直方图相结合的阈值法对荒漠草地植被覆盖度的提取效果较好,可以用于荒漠草地植被覆盖度的提取.     

基于照相法的荒漠草原植被覆盖度测定方法

荒漠草原植被稀疏、空间异质性与年季变化大。植被覆盖度是评价荒漠草原生态状况关键指标之一,但长期以来缺少快速、精准的测定方法。利用普通相机与ADC多光谱相机对地表拍照,结合监督分类、归一化差异指数法(以下简称NDI)提取地表植被覆盖度参数,并以20×20格网密度针刺法模拟结果为基准值,分析不同传感器及提取方式下的精度与适宜性。主要结果:(1)15×15格网密度针刺法可实现荒漠草原地表植被覆盖度高效、准确测定,当覆盖度高于60%时10×10格网密度最宜;(2)相比ADC多光谱照片及普通数码照片的监督分类,基于普通数码照片的NDI法能更准确提取荒漠草原植被低覆盖度,且较针刺法工作效率大幅提高。     

高山草地植被盖度对气候变暖和人类活动的响应

高寒草地对气候变化和人类活动的干扰响应敏感,高寒草地植被盖度估算是了解高寒草地植被状况和评估退化草地修复效果的一个重要方法。目前,地面盖度估算的主要方法是数码照相法,其中基于多光谱照片的盖度提取是比较准确和高效的方法。本研究采用多光谱相机对疏勒河上游不同演替阶段高寒草地、围栏封育和翻耕补播草甸生长季内的植被盖度进行了观测,研究了气候变暖和人类活动干扰下高寒草地植被盖度的变化,同时对估算多光谱照片盖度的阈值法做了改进。结果表明,1）新阈值法的估算结果与真实值之间无显著差异（P=0.219）,而旧阈值法与真实值之间差异极显著（P=0.001）;2）处于不同演替阶段的草地类型具有不同的植被盖度,沼泽草甸最高,各草地类型盖度均在7月达到最大;3）在生产力较低的多年冻土区,围栏封育对高寒草地修复效果在短期内不明显,翻耕补播甚至会导致植被盖度进一步降低。     

基于遥感技术的高寒草地覆盖度变化

摘要：本研究在像元二分法模型的基础上,利用改进归一化植被指数的植被覆盖度定量模型,采用MODIS卫星8天合成地表反射率数据产品（MOD09A1）,结合3S技术（GIS、RS、GPS）空间分析功能,以甘南州高寒草地为对象,分5个等级分别计算得到了甘南2000、2004和2008年的草地植被覆盖度,分析了甘南2000-2008年植被覆盖度变化的大致演变过程和趋势。结果表明,从2000-2008年,甘南植被覆盖度质量总体呈下降趋势,一级盖度植被退化比较严重,草地植被的演变情况主要由优等植被覆盖（一、二级）向低等植被覆盖（四、五级）演变。研究结果揭示了甘南州草地退化状况的动态趋势及严重程度,为相关政府和研究部门的草地管理决策提供了有效参考依据。     

基于数码相机的草地植被盖度测量方法对比研究

草地植被盖度是表征生态系统植被生长状况及环境质量的重要参数。在草地植物群落野外调查中,可以利用数码相机拍摄草地样方照片,而后在室内利用图像处理软件进行自动或半自动的植被盖度测量。随着移动智能设备(如iPhone/iPAD或各类Android Phone/PAD)的快速发展和普及,野外实时获取草地样方照片,同步计算草地植被盖度,并与有关遥感反演参数产品作校验对比分析,将成为未来地学移动测量和研究的重要方向。本研究在总结梳理既有利用数码相机识别植被盖度方法的基础上,设计了低覆盖、中低覆盖、中等覆盖、中高覆盖、高覆盖5种不同植被盖度情景,以及从早上6:00 到下午6:00、每隔2 h一次、全天共7次不同光照环境下的照相方案。继而以Photoshop人工勾勒和测算方法为基准,选择RGB阈值法、RGB决策树法、HSV判别法3种自动测量方法开展对比研究。测量结果的对比分析表明,草地盖度变化对RGB阈值法和HSV判别法的盖度识别精度无明显规律性影响,RGB决策树的盖度识别精度随着草地盖度的增加而增加;光照强度越强,RGB阈值法和HSV判别法对同一草地样方估算的盖度值越小,RGB决策树法估算的盖度值随光照强度的变化没有固定的规律。总体上讲,RGB阈值法和HSV判别法的识别精度较高,RGB决策树法误判率较高,但后者可以识别出非绿色的植物茎、花朵。最后提出了在现有绿色植被像元识别方法的基础上,结合边缘检查算法等图形像素的统计学特征分析方法,能进一步提高草地植被盖度测量的准确率。     

长江源区草地覆盖变化对土壤团聚体分布及稳定性的影响

为了明确长江源区草地覆盖变化对土壤结构的影响,本研究以高寒草甸和沼泽草甸2种草地为研究对象,分析了不同植被覆盖度下土壤团聚体分布和稳定性特征。结果显示：高寒草甸土壤团聚体主要分布在2~0.25 mm和0.25~0.053 mm粒径范围内,沼泽草甸土壤团聚体在不同粒径范围内分布相对均匀;高寒草甸土壤大于0.25 mm团聚体含量（WR_0.25）表现为高覆盖度草地高于中覆盖度草地,随土层加深,低覆盖度草地WR_0.25显著高于高覆盖度草地和中覆盖度草地,除表层0~10 cm土壤外,高覆盖度沼泽草甸土壤WR_0.25显著高于中覆盖度和低覆盖度土壤;不同植被覆盖度下高寒草甸表层0~10 cm土壤团聚体平均重量直径（Mean weight diameter,MWD）和几何平均直径（Geometric mean diameter,GMD）无显著差异,在10~20 cm和20~40 cm土层差异显著,表现为低覆盖度 > 高覆盖度 > 中覆盖度,沼泽草甸土壤团聚体MWD和GMD则随草地植被覆盖度降低和土层加深而逐渐减小;中覆盖度草地土壤团聚体破坏率（Aggregate destruction rate,PAD）最大,其稳定性最差;高覆盖度草地土壤团聚体稳定性最好;在不同土层深度上,高寒草甸土壤团聚体稳定性无显著差异,沼泽草甸土壤随土层加深团聚体稳定性变差。本研究结果表明WR_0.25和PAD能够更好地评价土壤团聚体稳定性,而沼泽草甸土壤团聚体对植被覆盖变化的响应更为敏感。本研究结果有利于加强植被覆盖变化背景下对长江源区土壤保护机制的认识。     

基于无人机大样方数据及国产卫星反演草地植被覆盖度方法研究

受开垦、采矿等人类活动影响,草原出现退化甚至沙化,监测其植被覆盖度对于揭示草地的分布状况与空间变化规律具有重要意义。本文以无人机大样方数据与国产高分一号（GF-1）数据作为数据源,结合野外同步数码相机获取的数据,应用支持向量机（Support vector machine,SVM）构建不同数据源之间的植被覆盖度反演模型（数码相片—无人机大样方数据植被覆盖度估算模型,无人机大样方数据—GF-1数据植被覆盖度估算模型）,探讨国产GF-1卫星结合无人机大样方估算草原植被覆盖度的方法。结果表明,基于SVM模型的GF-1数据结合无人机大样方计算的土壤调节植被指数（Soil-adjusted vegetation index,SAVI）具有较高的精度（判定系数R²=0.97,相对分析误差RPD=4.86,均方根误差RMSE=3.23）,因此基于无人机大样方数据结合GF-1数据可以准确、快速地反演草地覆盖度,利用这种方法可以估算整个草原的植被覆盖度。     

Precision,Repeatability, and Efficiency of Two Canopy-Cover Estimate Methods in Northern Great Plains Vegetation

Government agencies are subject to increasing public scrutiny of land management practices. Consequently, rigorous, yet efficient, monitoring protocols are needed to provide defensible quantitative data on the status and trends of rangeland vegetation. Rigor requires precise, repeatable measures, whereas efficiency requires the greatest possible information content for the amount of resources spent acquiring the information. We compared two methods—point frequency and visual estimate—of measuring canopy cover of individual plant species and groups of species (forbs vs. graminoids, native vs. nonnative) and plant species richness. These methods were compared in a variety of grassland vegetation types of the northern Great Plains for their precision, repeatability, and efficiency. Absolute precision of estimates was similar, but values generally differed between the two sampling methods. The point-frequency method yielded significantly higher values than the visual-estimate method for cover by individual species, graminoid cover, and total cover, and yielded significantly lower values for broadleaf (forb + shrub) cover and species richness. Differences in values derived by different sampling teams using the same method were similar between methods and within precision levels for many variables. Species richness and median species cover were the major exceptions; for these, the point-frequency method was far less repeatable. As performed in this study, the visual-estimate method required approximately twice the time as did the point-frequency method, but the former captured 55% more species. Overall, the visual-estimate method of measuring plant cover was more consistent among observers than anticipated, because of strong training, and captured considerably more species. However, its greater sampling time could reduce the number of samples and, therefore, reduce the statistical power of a sampling design if time is a limiting factor.     

Comparison of Two Sampling Methods for Quantifying Changes in Vegetation Composition Under Rangeland Development

Rapid vegetation sampling methods based on visual estimation are useful for monitoring changes in rangeland vegetation composition because large spatial and temporal scales are often involved and have limited sampling resources available. Here we compared two sampling methods in their ability to detect changes in vegetation composition following rangeland development: 1) species percent cover estimates within subplots (the percent cover [PC] method) and 2) rankings of relative biomass of the 10 most abundant species across the whole plot and the ratio of two of them (the visual ranking [VR] method). Both methods were applied on 30 experimental plots at year 26 of a long-term factorial trial of five soil fertility levels and three sheep grazing intensities. Multivariate statistical methods showed significant effects of experimental treatments (fertilizer level and sheep grazing intensity) and of vegetation sampling method (VR vs. PC) on vegetation composition. Importantly, we detected no significant interactions involving sampling method, indicating that the effect of sampling method was consistent across experimental treatments. Effects of fertilizer on vegetation composition were an order of magnitude greater than the effect of sampling method, whereas the latter was twice as important as the effect of grazing. Results were robust to differential weights given to relative abundances vs. compositional changes. Differences between methods were primarily driven by the PC method giving lower abundance estimates of one species, lupin (a hybrid of Lupinus polyphyllus Lindl.), relative to the VR method. Our results support the use of the VR method as a rapid yet powerful method for monitoring changes in vegetation composition under rangeland development.     

Mapping Total Vegetation Cover Across Western Rangelands With Moderate-Resolution Imaging Spectroradiometer Data

Remotely sensed observations of rangelands provide a synoptic view of vegetation condition unavailable from other means. Multiple satellite platforms in operation today (e.g. Landsat, moderate-resolution imaging spectroradiometer [MODIS]) offer opportunities for regional monitoring of rangelands. However, the spatial and temporal variability of rangelands pose challenges to consistent and accurate mapping of vegetation condition. For instance, soil properties can have a large impact on the reflectance registered at the satellite sensor. Additionally, senescent vegetation, which is often abundant on rangeland, is dynamic and its physical and photochemical properties can change rapidly along with moisture availability. Remote sensing has been successfully used to map local rangeland conditions. However, regional and frequently updated maps of vegetation cover in rangelands are not currently available. In this research, we compare ground measurements of total vegetation cover, including both green and senescent cover, to reflectance observed by the satellite and develop a robust method for estimating total vegetation canopy cover over diverse regions of the western United States. We test the effects of scaling from ground observations up to the Landsat 30-m scale, then to the MODIS 500-m scale, and quantify sources of noise. The soil-adjusted total vegetation index (SATVI) captures 55% of the variability in ground measured total vegetation cover from diverse sites in New Mexico, Arizona, Wyoming, and Nevada. Scaling from the Landsat to MODIS scale introduces noise and loss of spatial detail, but offers inexpensive and frequent observations and the ability to track trends in cover over large regions.     

祁连山高寒草原土壤水分与植被盖度的关系

为了揭示高寒草原土壤水分与植被盖度的关系,提高植被盖度与土壤水分快测的精度,本研究在祁连山北麓中段高寒草原—甘肃马鹿冬季牧场,设置放牧率分别为1.00、1.45、2.45、3.45、4.85、6.90 AUM·hm^-2的6个样地,通过逐月测定生长季植被盖度和土壤水分,结合Landsat5多光谱影像与SPOT2全色影像融合而成的SPOT-TM影像,分析土壤水分与植被盖度、归一化植被指数(NDVI)的关系,并在此基础上,确定植被盖度随土壤水分的变化阶段。除醉马草、银灰旋花、乳白黄耆外,其他种群盖度对表层土壤水分均响应敏感,在3%~7%至17%~26%的土壤水分范围内,种群盖度随土壤水分的增加呈反比例函数递增趋势,且增幅逐渐缩小;群落盖度对表层土壤水分的响应敏感,在1.1%~10.0%的土壤水分范围内,群落盖度随土壤水分的增加亦呈反比例函数递增趋势,并逐渐趋于饱和。研究结果对探讨干旱半干旱草原植被物种、种群、群落的抗旱性、水分利用效率具有一定意义,为草原植被和土壤水分管理提供科学依据。     

A Method For Landscape-Scale Vegetation Assessment: Application to Great Basin Rangeland Ecosystems

The growth of landscape-scale land management necessitates the development of methods for large-scale vegetation assessment. Field data collection and analysis methods used to assess ecological condition for the 47 165-h North Spring Valley watershed are presented. Vegetation cover data were collected in a stratified random design within 6 Great Basin vegetation types, and the probability of detecting change in native herbaceous cover was calculated using power analyses. Methods for using these quantitative assessment data are presented to calculate a departure index based on reference condition information from LANDFIRE (an interagency effort to map and model fire regimes and other biophysical characteristics at a mid-scale for the entire United States) Biophysical Setting models for the mountain big sagebrush (Artemisia tridentata Nutt. subsp. vaseyana [Rydb.] Beetle) vegetation type. For mountain big sagebrush in the North Spring Valley landscape, we found that the earliest successional classes were underrepresented and that mountain big sagebrush moderately invaded by conifers was more abundant than predicted by the LANDFIRE reference based on the historic range of variability. Classes that were most similar to the reference were mountain big sagebrush with the highest conifer cover and late development mountain big sagebrush with perennial grasses. Overall, results suggested that restoration or approximation of the historic fire regime is needed. This method provides a cost-effective procedure to assess important indicators, including native herbaceous cover, extent of woody encroachment, and ground cover. However, the method lacks the spatial information that would allow managers to comprehensively assess spatial patterns of vegetation condition across the mosaics that occur within each major vegetation type. The development of a method that integrates field measurements of key indicators with remotely sensed data is the next critical need for landscape-scale assessment.     

Comparison of Unmanned Aerial Vehicle Platforms for Assessing Vegetation Cover in Sagebrush Steppe Ecosystems

In this study, the use of unmanned aerial vehicles (UAVs) as a quick and safe method for monitoring biotic resources was evaluated. Vegetation cover and the amount of bare ground are important factors in understanding the sustainability of many ecosystems. Methods that improve speed and cost efficiency could greatly improve how biotic resources are monitored on western lands. Sagebrush steppe ecosystems provide important habitat for a variety of species including sage grouse and pygmy rabbit. Improved methods of monitoring these habitats are needed because not enough resource specialists or funds are available for comprehensive on-the-ground evaluations. In this project, two UAV platforms, fixed-wing and helicopter, were used to collect still-frame imagery to assess vegetation cover in sagebrush steppe ecosystems. This paper discusses the process for collecting and analyzing imagery from the UAVs to 1) estimate percentage of cover for six different vegetation types (shrub, dead shrub, grass, forb, litter, and bare ground) and 2) locate sage grouse using representative decoys. The field plots were located on the Idaho National Laboratory site west of Idaho Falls, Idaho, in areas with varying amounts and types of vegetation cover. A software program called SamplePoint was used along with visual inspection to evaluate percentage of cover for the six cover types. Results were compared against standard field measurements to assess accuracy. The comparison of fixed-wing and helicopter UAV technology against field estimates shows good agreement for the measurement of bare ground. This study shows that if a high degree of detail and data accuracy is desired, then a helicopter UAV may be a good platform to use. If the data collection objective is to assess broad-scale landscape level changes, then the collection of imagery with a fixed-wing system is probably more appropriate.     

Performance of Quantitative Vegetation Sampling Methods Across Gradients of Cover in Great Basin Plant Communities

Resource managers and scientists need efficient, reliable methods for quantifying vegetation to conduct basic research, evaluate land management actions, and monitor trends in habitat conditions. We examined three methods for quantifying vegetation in 1-ha plots among different plant communities in the northern Great Basin: photography-based grid-point intercept (GPI), line-point intercept (LPI), and point-quarter (PQ). We also evaluated each method for within-plot subsampling adequacy and effort requirements relative to information gain. We found that, for most functional groups, percent cover measurements collected with the use of LPI, GPI, and PQ methods were strongly correlated. These correlations were even stronger when we used data from the upper canopy only (i.e., top “hit” of pin flags) in LPI to estimate cover. PQ was best at quantifying cover of sparse plants such as shrubs in early successional habitats. As cover of a given functional group decreased within plots, the variance of the cover estimate increased substantially, which required more subsamples per plot (i.e., transect lines, quadrats) to achieve reliable precision. For GPI, we found that that six–nine quadrats per hectare were sufficient to characterize the vegetation in most of the plant communities sampled. All three methods reasonably characterized the vegetation in our plots, and each has advantages depending on characteristics of the vegetation, such as cover or heterogeneity, study goals, precision of measurements required, and efficiency needed.     

Using Unmanned Helicopters to Assess Vegetation Cover in Sagebrush Steppe Ecosystems

Evaluating vegetation cover is an important factor in understanding the sustainability of many ecosystems. Remote sensing methods with sufficient accuracy could dramatically alter how biotic resources are monitored on both public and private lands. Idaho National Laboratory (INL), in conjunction with the University of Idaho, evaluated whether unmanned aerial vehicles (UAVs) are sufficiently accurate and more efficient than the point-frame field method for monitoring vegetative cover and bare ground in sagebrush steppe ecosystems. These values are of interest to land managers because typically there are limited natural resource scientists and funding for comprehensive ground evaluations. In this project, unmanned helicopters were used to collect still-frame imagery to determine vegetation cover during June and July 2005. The images were used to estimate percent cover for six vegetative cover classes (shrub, dead shrub, grass, forbs, litter, and bare ground). Field plots used to collect imagery and on-the-ground measurements were located on the INL site west of Idaho Falls, Idaho. Ocular assessments of digital imagery were performed using SamplePoint, and the results were compared with field measurements collected using a point-frame method. The helicopter imagery evaluation showed a high degree of agreement with field cover class values for grass, litter, and bare ground and reasonable agreement for dead shrubs. Shrub cover was often overestimated, and forbs were generally underestimated. The helicopter method took 45% less time than the field method. This study demonstrates that UAV technology provides a viable method for monitoring selective types of cover on rangelands and could save time and resources.     

四种覆土处理对高寒煤矿区排土场渣山植被恢复的影响

通过覆土措施对高寒煤矿区排土场渣山表层进行土壤重构,加速植被生长是高寒矿区植被恢复的关键技术问题。本研究通过在青海木里煤田圣雄煤矿建立试验小区,比较覆土0 (对照)、覆土5 (FT-5)、覆土10 (FT-10)和覆土15 cm(FT-15)4种不同覆土处理对高寒矿区排土场渣山植被恢复的影响,研究高寒矿区排土场渣山表层土壤重构技术,为高寒矿区植被恢复提供科学依据。采用由垂穗披碱草、中华羊茅、青海冷地早熟禾、青海草地早熟禾、星星草5种牧草混播种植方案开展研究。结果表明,垂穗披碱草盖度在所有播种牧草中最大,是混播优势牧草,覆土处理小区垂穗披碱草高度逐年增加。4种覆土处理植被3年平均总盖度排序结果为覆土15 cm(74.6%)>覆土10 cm(70.8%)>覆土5 cm(64.3%)>对照(59.8%)。随处理时间延长,各处理间牧草植株密度差异显著性越来越大。经过3年试验,对照小区牧草产量同覆土小区存在极显著差异(P<0.01)。产量最高的是覆土10 cm小区,达到每小区(453.6±38.4) g·m^-2,是对照区域的1.94倍。通过相关性分析发现,牧草高度、密度分别与速效氮含量呈显著正相关关系(r=0.578和r=0.619,P<0.05),产量也呈极显著正相关(r=0.839,P<0.01),有机质含量同小区产量显著相关(r=0.592,P<0.05)。经推算,覆土5、10、15 cm成本分别为20000、40000 和60000元·hm^-2。覆土可以有效促进植物生长,提高植被的高度、盖度和密度,覆土10 cm是高寒煤矿区相对比较理想的覆土重构措施。