基于决策树方法的云南省森林分类研究

森林分类对于理解森林生态系统结构和功能具有重要意义。由于云南省地形和森林类型复杂,首先按云南省的16个行政区划将全省Landsat TM影像分为对应的16个区域。以TM波段1~5和7,以及由植被指数、缨帽变换、主成分变换、DEM组成的18个变量组,统计训练样本光谱值均值变化和光谱值与频率间的关系。利用交点计算公式计算类间最佳分类界点进而建立决策树,逐一分离各区的所有森林类型,将分类结果合并得到云南省阔叶林、针叶林和针阔混交林类型分布图。最后将分类结果与监督分类中的最大似然比法的分类结果进行对比。结果表明:监督分类的总体分类精度为74.39%,Kappa系数为0.63,决策树方法的总体分类精度为86.61%,Kappa系数为0.80,说明决策树方法可以提取高精度的云南省森林类型,进而为该区域森林叶面积指数和生物量反演等研究提供基础数据支持。     

基于背包激光雷达的单木材积估测方法研究

背包式激光雷达(Backpack Laser Scanning,BLS)在森林资源调查中具有很大的应用潜力,但在复杂地表情景下,单木材积和林分蓄积量提取精度存在较大不确定性。以广西高峰林场为研究区,利用随机森林方法,基于BLS点云数据对单木材积和样地蓄积量进行估测。首先,对BLS点云进行单木分割,提取单木胸径(DBH)、树高(Htree)、冠幅直径(CD)、冠幅面积(CA)、冠幅体积(CV)、郁闭度(CC)、间隙率(GF)和叶面积指数(LAI)共8个特征参数,并计算56个分层高度指标(高度百分比、累积高度百分比、变异系数、冠层起伏率等)。然后,通过随机森林算法构建单木材积估测模型,并对比各种参数组合的预测精度。得到结果：(1)仅用8个单木结构特征参数进行建模,估测精度为：R²=0.83、RMSE=0.097 m³;(2)加入分层高度指标的模型估测精度有所提升：R²=0.87、RMSE=0.087 m³;(3)通过Boruta算法进行变量筛选,输入参数从64个减少至52个,估测精度差异不大：R     

激光点云深度学习的树种识别研究北大核心CSCD

针对激光雷达林业树种分类难以直接使用点云数据的问题,使用基于点云深度学习方法进行树种识别并提出PointNet-GS模型,无需将点云转为三维体素或二维图像,避免数据类型转换造成的特征丢失。以河北省塞罕坝机械林场的落叶松和白桦两个树种为研究对象。首先,将获取的点云数据进行数据预处理、单木分割,提取分割效果较好的单木作为样本;其次,将单木提取的样本进行几何下采样处理,保留更多局部特征便于网络模型学习;最后,将下采样处理的样本输入深度学习模型的网络,自动提取其高维特征进行学习,实现树种分类。实验结果表明,PointNet-GS树种分类精度达89.3%,Kappa系数为0.785,效果优于原始PointNet模型。     

基于GEE云平台和Sentinel-2数据的普洱市森林覆盖制图

快速准确获取森林的空间分布对评估森林资源和生态环境状况具有重要的意义。以云南省普洱市为研究区,基于Google Earth Engine(GEE)平台和Sentinel-2影像数据,结合实地调查数据、机载遥感数据及地形辅助数据,提取影像的光谱特征、纹理特征以及地形特征,通过特征筛选,得到适合森林分类的最优特征数据集。结合简单线性非迭代聚类（SimpleNon-Iterative Clustering,SNIC）超像素分割算法,探究不同分类方法、特征变量等因素对分类精度的影响。结果表明：面向对象分类方法的分类精度要优于基于像元分类方法,分类总体精度为88.21%,Kappa系数为0.87,可以较为准确地对普洱市进行森林覆盖制图。面向对象方法可以有效减轻“椒盐现象”,特征优选避免了冗余信息对分类结果的影响,有效提高了分类效率。GEE平台与面向对象方法结合可以提供大区域、高精度的森林覆盖遥感快速制图。     

基于遥感光谱和空间变量随机森林的黄河三角洲刺槐林健康等级分类

对刺槐林健康状况进行准确分类制图,是进行刺槐林健康状况评估与生态修复的前提。以高分辨率IKONOS影像、基于影像提取的不同窗口、不同灰度共生矩阵纹理信息以及反映局部空间自相关的Local Getis-Ord Gi(Getis统计量)为数据源,结合实测生态样方数据,利用多决策树的组合分类模型随机森林(RF)对刺槐林健康进行分级,对6种方法的分类精度进行了比较且对分类变量的重要性进行了排序。结果显示:19m×19m是最佳纹理计算窗口;灰度共生矩阵均值是最优纹理变量;基于波段4计算的Getis统计量对RF分类具有最重要的作用;较之利用全部光谱、纹理和Getis统计量的80个波段/变量,利用前向选择得到的前16个重要性变量进行RF分类,获得了最高的分类精度(总精度为93.14%,Kappa系数为0.894)。研究证实了从高分影像提取的空间特征信息有助于提高对具有规则分布格局的人工刺槐林健康等级的分类精度;前向选择方法可以利用较少的预测变量获得较高的分类精度。     

基于不同特征的随机森林极化SAR图像分类

近些年,利用计算机对极化SAR图像进行分类逐渐成为遥感领域的一个研究热点.本文采用全极化SAR数据,利用不同的特征提取算法提取特征,并基于随机森林模型最终实现对江苏沿海滩涂的分类.首先采用H/α和Freeman两种分解算法提取极化特征参数,采用灰度共生矩阵提取纹理特征参数;然后将提取的所有特征进行不同的组合,构成不同的特征集;最后采用随机森林模型对不同特征集合进行分类和精度评估.结果表明仅用纹理特征对沿海滩涂进行分类时效果较差;利用极化分解提取出的散射特征进行分类的结果要优于矩阵元素特征的分类结果;综合了极化散射特征和纹理特征的组合方式在沿海滩涂的分类中可以取得最优的分类结果,总体精度和Kappa系数可以达到94.44%和0.9305,表明极化SAR图像中蕴含的不同方面的特征在分类中具有一定的互补性.     

一种高分辨率遥感图像单木树冠信息提取方法

单木树冠信息是森林管理和相关科学研究的基础数据。为解决面向对象方法在树冠相互连接时无法有效提取单木树冠信息的问题,提出了一种基于面向对象技术和水文分析的高分辨率遥感图像单木树冠信息提取方法。方法首先对图像进行预处理,然后利用面向对象方法从融合图像获取树冠的分布范围,利用水文分析技术从全色波段图像获取单木树冠的潜在分布范围,最后将提取的数据进行叠加相交处理,完成单木位置信息提取和单木树冠描绘。实验结果表明,所提出的方法能有效提取单木位置和描绘单木树冠;采用修改的遥感分类精度评价指标进行的精度评价结果为,单木探测总体精度为87.63%,单木树冠描绘总体精度为84.89%。     

基于高分二号遥感影像的树种分类方法

为推广国产高分数据在森林树种分类方面的应用,以北京市延庆区八达岭国家森林公园主要区域的6期高分二号影像为数据源,在分层分类的基础上,利用支持向量机递归特征消除、C5.0决策树、FSO 3种特征优选方法,从4种特征维度下实现面向对象的支持向量机和随机森林的森林树种分类,最终取得总体精度平均为83.65%,特定树种生产者精度介于93.75%（山杏）和38.10%（刺槐）之间,特定树种用户精度介于100%（华北落叶松）和44.74%（榆树）之间的良好结果。结果表明：C5.0决策树耗时最短（0.01 h）且其所选特征应用于分类总体精度最高（86.90%）;在不同特征维度下支持向量机分类的总体精度比随机森林平均高出3.28%;支持向量机和随机森林均对特征维度不敏感,但良好的特征优选结果仍会对支持向量机的分类效率（最高提升86.98%）和随机森林的分类精度（最高提升9.22%）产生较大影响。     

遥感图像森林类型小波纹理的SVM法分类

利用遥感图像对森林类型进行分类是大面积地调查、监测、分析森林资源的快速与经济的方法,但由于不同森林的光谱特征非常相近而较难准确分类。因此,在GPS数据和高分辨率遥感图像的支持下,对水源林Landsat TM遥感图像用窗口法获得阔叶林、针叶林和竹林样本图像,然后计算其小波分解后小波系数的l1范数纹理测度构成分类特征向量,利用支持向量基SVM进行分类。结果表明,利用SVM对图像中阔叶林、针叶林和竹林分类平均精度在80%以上,可较准确地识别森林类型,图像总体分类精度达到90.2%,Kappa系数0.77,均比利用小波纹理特征的神经网络法和最大似然法有所提高,森林分类错误产生的主要原因是混交林造成两类森林间存在交集。该方法可以较有效地提高遥感图像森林类型的分类精度。     

基于高分二号遥感影像的树种分类方法

为推广国产高分数据在森林树种分类方面的应用，以北京市延庆区八达岭国家森林公园主要区域的6期高分二号影像为数据源，在分层分类的基础上，利用支持向量机递归特征消除、C5.0决策树、FSO 3种特征优选方法，从4种特征维度下实现面向对象的支持向量机和随机森林的森林树种分类，最终取得总体精度平均为83.65%，特定树种生产者精度介于93.75%（山杏）和38.10%（刺槐）之间，特定树种用户精度介于100%（华北落叶松）和44.74%（榆树）之间的良好结果。结果表明：C5.0决策树耗时最短（0.01 h）且其所选特征应用于分类总体精度最高（86.90%）；在不同特征维度下支持向量机分类的总体精度比随机森林平均高出3.28%；支持向量机和随机森林均对特征维度不敏感，但良好的特征优选结果仍会对支持向量机的分类效率（最高提升86.98%）和随机森林的分类精度（最高提升9.22%）产生较大影响。     

遥感图像林型纹理特征的ICA与SVM分类

遥感图像纹理特征是光谱相近林型准确分类的有效方法,然而其带来分类特征向量维数增加和计算量增大。因此,对南方山区林地TM图像进行独立成分分析ICA降维,通过计算灰度共生矩阵获取纹理特征,使用SVM分类,研究林地类型的快速分类方法。结果表明,ICA与SVM法利用遥感图像纹理特征可较准确地实现林地类型分类,分类总精度、Kappa系数分别为85.4%、0.73,均高于SVM法、BP神经网络法、最大似然法、最小距离法;其对阔叶林、针叶林、竹林的分类精度依次为78.2%、80.1%、84.3%,误识率主要是由于混交林而造成两类林地之间存在交集,易出现的针阔混交林使得阔叶林、针叶林的分类精度低于竹林。     

基于无人机高光谱影像和机器学习的红树林树种精细分类

利用海南省文昌市清澜港红树林保护区的无人机高光谱影像,采用递归特征消除的随机森林算法(Recursive Feature Elimination-Random Forest,RFE-RF)优选植被光谱特征和纹理特征,通过机器学习中的随机森林(Random Forest,RF)和支持向量机(Support Vector ...     

Airborne hyperspectral discrimination of tree species with different ages using discrete wavelet transform

In this article, the capability of discrete wavelet transform (DWT) to discriminate tree species with different ages using airborne hyperspectral remote sensing is investigated. The performance of DWT is compared against commonly used traditional methods, i.e. original reflectance and first and second derivatives. The hyperspectral data are obtained from Thetford forest of the UK, which contains Corsican and Scots pines with different ages and broadleaved tree species. The discrimination is performed by employing three different spectral measurement techniques (SMTs) including Spectral Angle Mapper (SAM), Spectral Information Divergence (SID), and a combination of SAM and SID. Five different mother wavelets with a total of 50 different orders are tested. The wavelet detail coefficient (CD) from each decomposition level and combination of all CDs plus the approximation coefficient from the final decomposition level (C-All) are extracted from each mother wavelet. The results show the superiority of DWT against the reflectance and derivatives for all the three SMTs. In DWT, C-All provided the highest discrimination accuracy compared to other coefficients. An overall accuracy difference of about 20–30% is observed between the finest coefficient and C-All. Amongst the SMTs, SID provided the highest accuracy, while SAM showed the lowest accuracy. Using DWT in combination with SID, an overall accuracy up to around 71.4% is obtained, which is around 13.5%, 14.7%, and 27% higher than the accuracies achieved with reflectance and first and second derivatives, respectively.     

基于深度信念网络的高光谱影像森林类型识别

森林类型分类对森林生态系统管理起重要作用,高光谱影像由于波段多,传统方法先对其进行特征选择或特征提取进行降维处理,再进行图像分类,一定程度影响森林类型识别精度.深度信念网络是一种半监督学习方法,可将高光谱所有波段作为深度信念网络的输入,从而避免降维处理.论文利用深度信念网络对泉州市德化县西部8个乡镇进行森林类型识别研究.基于HJ/1A高光谱图像与二类调查数据,利用Python语言实现高光谱影像森林类型分类,讨论了网络深度和隐藏层单元数对总体精度与Kappa系数的影响.实验结果表明:层数为3,每层节点数为256的网络结构对森林类型识别效果最好,总体精度达85.8%,系数为0.785,好于支持向量机的分类结果.     

Comparison of Machine Learning Methods for Land Use/Land Cover Classification in the Complicated Terrain Regions

Aiming at the characteristics of varied and complex geomorphic types,crisscross network of ravines and broken terrain in high altitude complicated terrain regions,it is very important to study and find the rapid and effective land use/land cover classification method for obtaining and timely updating of land use information.Taking the Huangshui river basin located in the transitional zone between the Loess Plateau and the Qinghai-Tibet Plateau as acasestudy area,the objective of this study is to explore a kind of effective information extraction method from comparison of four kinds machine learning methods for complicated terrain regions.based on Landsat 8 OLI satellite data,DEM and combined with various thematic features,on the basis of geographical division of the study area,artificial neural network,decision tree,support vector machine and random forest four machine learning methods for land use information extraction were used to obtain land use data,and confusion matrix was constructed to evaluate classification accuracy.The results showed that the classification accuracies of random forest and decision tree are obviously higher than those of support vector machine and artificial neural network.The random forest method has the highest classification accuracy,the overall classification accuracy is 85.65%,the Kappa coefficient is 0.84.based on the above classification,Random forest classification method was chose to further classify Landsat 8 fusion datafrom panchromatic 15 meter and multispectral 30 meter image,the overall classification accuracy is 86.49% and the Kappa coefficient is 0.85.This indicated that the random forest classification method can obtain higher classification efficiency while ensuring the classification accuracy.It is very effective for the extraction of land use information in complicated terrain regions.Data fusion can improve the classification accuracy to a certain extent.     

Segmented canonical discriminant analysis of in situ hyperspectral data for identifying 13 urban tree species

A total of 458 in situ hyperspectral data were collected from 13 urban tree species in the City of Tampa, FL, USA using a spectrometer. The 13 species include 11 broadleaf and two conifer species. Three different techniques, segmented canonical discriminant analysis (CDA), segmented principal component analysis (PCA) and segmented stepwise discriminate analysis (SDA), were applied and compared for dimension reduction and feature extraction. With each of the three techniques, 10 features were extracted or selected from four spectral regions, visible (VIS: 1412–1797 nm), near-infrared (NIR: 707–1352 nm), mid-infrared 1 (MIR1: 1412–1797 nm) and mid-infrared 2 (MIR2: 1942–2400 nm), and used to discriminate the 13 urban tree species with a linear discriminate analysis (LDA) method. The cross-validation results, based on training samples that were used in the feature reduction step, and the results calculated from the test samples were used for evaluating the ability of the in situ hyperspectral data and performance of the segmented CDA, PCA and SDA to identify the 13 tree species. The experimental results indicate that a satisfactory discrimination of the 13 tree species was achieved using the segmented CDA technique (average accuracy (AA) = 96%, overall accuracy (OAA) = 96% and kappa = 0.958 from the cross-validation results; AA = 90%, OAA = 90% and kappa = 0.896 from the test samples) compared to the segmented PCA and SDA techniques, respectively (AA = 76% and 86%, OAA = 78% and 87%, and kappa = 0.763 and 0.857 from the cross-validation results; AA = 79% and 88%, OAA = 80% and 89%, and kappa = 0.782 and 0.879 from the test samples). In this study, the segmented CDA transformation is effective for dimension reduction and feature extraction for species discrimination with a relatively limited number of training samples. It outperformed the segmented PCA and SDA methods and produced the highest accuracies. The NIR and MIR1 regions have greater power for identifying the 13 species compared to the VIS and MIR2 spectral regions. The results indicate that CDA or segmented CDA could be applied broadly in mapping forest cover types, species identification and/or other land use/land cover classification practices with hyperspectral remote sensing data.     

Assessing the utility of airborne hyperspectral and LiDAR data for species distribution mapping in the coastal Pacific Northwest, Canada

To effectively manage forested ecosystems an accurate characterization of species distribution is required. In this study we assess the utility of hyperspectral Airborne Imaging Spectrometer for Applications (AISA) imagery and small footprint discrete return Light Detection and Ranging (LiDAR) data for mapping 11 tree species in and around the Gulf Islands National Park Reserve, in coastal South-western Canada. Using hyperspectral imagery yielded producer's and user's accuracies for most species ranging from > 52-95.4 and > 63-87.8%, respectively. For species dominated by definable growth stages, pixel-level fusion of hyperspectral imagery with LiDAR-derived height and volumetric canopy profile data increased both producer's (+ 5.1-11.6%) and user's (+ 8.4-18.8%) accuracies. McNemar's tests confirmed that improvements in overall accuracies associated with the inclusion of LiDAR-derived structural information were statistically significant (p < 0.05). This methodology establishes a specific framework for mapping key species with greater detail and accuracy then is possible using conventional approaches (i.e., aerial photograph interpretation), or either technology on its own. Furthermore, in the study area, acquisition and processing costs were lower than a conventional aerial photograph interpretation campaign, making hyperspectral/LiDAR fusion a viable replacement technology.     

Multispectral and LiDAR data fusion for fuel type mapping using Support Vector Machine and decision rules

This paper presents a method for mapping fuel types using LiDAR and multispectral data. A two-phase classification method is proposed to discriminate the fuel classes of the Prometheus classification system, which is adapted to the ecological characteristics of the European Mediterranean basin. The first step mapped the main fuel groups, namely grass, shrub and tree, as well as non-fuel classes. This phase was carried out using a Support Vector Machine (SVM) classification combining LiDAR and multispectral data. The overall accuracy of this classification was 92.8% with a kappa coefficient of 0.9. The second phase of the proposed method focused on discriminating additional fuel categories based on vertical information provided by the LiDAR measurements. Decision rules were applied to the output of the SVM classification based on the mean height of LiDAR returns and the vertical distribution of fuels, described by the relative LiDAR point density in different height intervals. The final fuel type classification yielded an overall accuracy of 88.24% with a kappa coefficient of 0.86. Some confusion was observed between fuel types 7 (dense tree cover presenting vertical continuity with understory vegetation) and 5 (trees with less than 30% of shrub cover) in some areas covered by Holm oak, which showed low LiDAR pulses penetration so that the understory vegetation was not correctly sampled.     

Improving the classification of six evergreen subtropical tree species with multi-season data from leaf spectra simulated to WorldView-2 and RapidEye

Remote sensing offers a feasible means to monitor tree species at a regional level where species distribution and composition is affected by the impacts of global change. Furthermore, the temporal resolution of space-borne multispectral sensors offers the ability to combine phenologically important phases for the optimization of tree species classification. In this study, we determined whether multi-seasonal leaf-level spectral data (winter, spring, summer, and autumn) improved the classification of six evergreen tree species in the subtropical forest region of South Africa when compared to a single season, for hyperspectral data, and reflectance data simulated to the WorldView-2 (WV2) and RapidEye (RE) sensors. Classification accuracies of the test data were assessed using a Partial Least Square Random Forest algorithm. The accuracies were compared between single seasons and multi-season classification and across seasons using analysis of variance and post-hoc Tukey Honest Significant Difference tests. The average overall accuracy (OA) of the leaf-level hyperspectral data ranged from a minimum of 90 ± 3.5% in winter to a maximum of 92 ± 2.7% in summer, outperforming the simulated reflectance data for the WV2 and RE sensors with an average OA of between 8 and 10 percentage points (p < 0.02, Bonferroni corrected). The use of data from multiple seasons increased the average OA and decreased the number of species pair confusions for the simulated multispectral classifications. The producer’s and user’s accuracies of the hyperspectral classification were >82% and showed no significant change using multi-season data. Multiple seasons may therefore be beneficial to multispectral sensors with ≤8 bands, yet remains to be tested at canopy level, for other species and climatic regions.