Comparative performance of convolutional neural network,weighted and conventional support vector machine and random forest for classifying tree species using hyperspectral and photogrammetric data

ABSTRACT

The classification of tree species can significantly benefit from high spatial and spectral information acquired by unmanned aerial vehicles (UAVs) associated with advanced classification methods. This study investigated the following topics concerning the classification of 16 tree species in two subtropical forest fragments of Southern Brazil: i) the potential integration of UAV-borne hyperspectral images with 3D information derived from their photogrammetric point cloud (PPC); ii) the performance of two machine learning methods (support vector machine – SVM and random forest – RF) when employing different datasets at a pixel and individual tree crown (ITC) levels; iii) the potential of two methods for dealing with the imbalanced sample set problem: a new weighted SVM (wSVM) approach, which attributes different weights to each sample and class, and a deep learning classifier (convolutional neural network – CNN), associated with a previous step to balance the sample set; and finally, iv) the potential of this last classifier for tree species classification as compared to the above mentioned machine learning methods. Results showed that the inclusion of the PPC features to the hyperspectral data provided a great accuracy increase in tree species classification results when conventional machine learning methods were applied, between 13 and 17% depending on the classifier and the study area characteristics. When using the PPC features and the canopy height model (CHM), associated with the majority vote (MV) rule, the SVM, wSVM and RF classifiers reached accuracies similar to the CNN, which outperformed these classifiers for both areas when considering the pixel-based classifications (overall accuracy of 84.4% in Area 1, and 74.95% in Area 2). The CNN was between 22% and 26% more accurate than the SVM and RF when only the hyperspectral bands were employed. The wSVM provided a slight increase in accuracy not only for some lesser represented classes, but also some major classes in Area 2. While conventional machine learning methods are faster, they demonstrated to be less stable to changes in datasets, depending on prior segmentation and hand-engineered features to reach similar accuracies to those attained by the CNN. To date, CNNs have been barely explored for the classification of tree species, and CNN-based classifications in the literature have not dealt with hyperspectral data specifically focusing on tropical environments. This paper thus presents innovative strategies for classifying tree species in subtropical forest areas at a refined legend level, integrating UAV-borne 2D hyperspectral and 3D photogrammetric data and relying on both deep and conventional machine learning approaches.     

广州华侨新村低密度住宅区庭院的树木和景观研究*

对广州华侨新村低密度住宅区庭院树木的组成、结构、空间格局、树木与生境的关系,以及私人庭院树木与公共绿地的不同特点进行了研究。结果表明:华侨新村庭院树木的种类多样性很高,共有114个树种。与广州城市行道树相比,华侨新村的乡土树种和果树比例较高。在调查的树木中84.75%的树木没有结构损伤现象,只有1.5%的树木生长较差。私人庭院内的树种较为分散,有45种树木为每个庭院仅一棵的单生种,有67.1%的庭院其树种在5种或以上。树木在街道的空间分布则有相对集中的趋势。庭院的大小是影响树木生长的重要条件,在有限空间的庭院内种植过多的树木是造成树木和环境关系紧张的重要原因。     

钻探工程人身伤害分析

利用数理统计方法，对某地勘队１５年中出现的钻探人身伤害，进行了归纳和分析，探讨了钻探伤害日内出现的规律，伤害与钻探工龄之间的关系以及钻探伤害的人体部位与伤害形式等。     

基于最小生成树的Steiner最小树生成算法

提出了基于最小生成树的Steiner最小树的生成算法,分析了该算法的时间复杂性为O(nlogn).     

GIS辅助下的基于数据挖掘的林型遥感分类方法研究

在植被分类领域,各类植被光谱特征非常相近,纹理特征比较单一,并且受物候影响较大,使得影像上林地信息的分类识别的成为一个较复杂的过程。本文利用GIS平台构建由影像、DEM等组成的多元空间数据库,将数据挖掘的方法引入林型遥感分类中,充分利用了现有的空间数据资源,减少了同物异谱、异物同谱对林型遥感分类的影响,增强了分类知识获取的自动化和智能化程度。实验结果表明,该方法具有简化知识获取过程和提高分类精度等优点。     

基于区间树的等高线提取算法研究(英文)

This paper proposes a new algorithm for determining the starting points of contour lines. The new algorithm is based on the interval tree. The result improves the algorithm＇s efficiency remarkably. Further, a new strategy is designed to constrain the direction of threading and the resulting contour bears more meaningful information.     

Annual growth zones in stems of Hypericum irazuense (Guttiferae) in the Costa Rican páramos

The high peaks of the Cordillera de Talamanca in Costa Rica support shrub- and grass-dominated páramo ecosystems that experience stand-replacing wildfires. The dry season that facilitates these fires results in dormancy in plant growth and provides an opportunity to use dendrochronological analyses to determine ages of plants in burn sites to support studies of fire history and postfire vegetation recovery. This study investigates the formation of annual growth zones in stems of the common shrub Hypericum irazuense. Unlike other páramo shrubs, H. irazuense rarely resprouts following fire, instead recovering through seedling recruitment following seed dispersal from the unburned periphery. Laboratory analysis of 19 prepared cross sections from 15 stems shows that samples of H. irazuense from burned areas can provide an estimate of the minimum number of years since the previous fire, supporting earlier work based on field examination of stems. Including a time lag for seedling recruitment or resprouting refines that estimate. Counts of growth zones in most sections coincided with dates of the last known fires. The presence of annual growth zones in H. irazuense places the species within a relatively small group of woody Neotropical species for which annual rings or growth zones have been demonstrated.     

榆树年轮记录的浑善达克沙地春季平均最高气温

基于采自内蒙古正蓝旗北部浑善达克沙地的榆树树芯及圆盘建立了区域年表,其与浑善达克沙地4—5月平均最高气温存在显著的负相关性（r=-0.72）,根据二者的相关性建立回归方程（方差解释量为51.35%）,稳定可靠,并据此重建了该地区1766—2011年4—5月平均最高气温,共246 a。重建序列中,气温显著偏高的时段有5个,显著偏低的时段有6个。采用多窗谱分析检测序列周期性,存在2~3.5 a和33~51.28 a周期。重建序列与周围地区干旱指数重建序列波动趋势具有一致性,且周围地区4—5月平均最高气温与重建的该地区4—5月平均最高气温间相关系数均高于0.6。     

Long-term changes in the tree radial growth and intrinsic water-use efficiency of Chuanxi spruce (<Emphasis Type="Italic">Picea likiangensis var. balfouriana</Emphasis>) in southwestern China

Elevated CO₂ level in the atmosphere is expected to improve the tree growth rates and intrinsic water-use efficiency (iWUE). Although current results inferring from tree rings found the tree growth decline in water-limited area, it is still unclear whether spruce trees in humid southwestern China benefit from the increasing CO₂. In this study, tree-ring width data were used to investigate the tree radial growth rate of Chuanxi spruce (Picea likiangensis var. balfouriana). Moreover, combining with the tree-ring carbon isotope date, we analyzed the physiological responses of Chuanxi spruce to rising CO₂ concentrations in the atmosphere (Ca) associated with climatic change in southwestern China. From 1851 to 2009, iWUE of Chuanxi spruce rose by approximately 30.4%, and the ratio of atmospheric CO₂ to leaf intercellular CO₂ concentration (Ci/Ca) showed no significant trend in the study area. The result suggested that Chuanxi spruce used an active response strategy when Ca was significantly increased. iWUE showed a significant increasing trend in parallel with tree radial growth, indicating that the increasing iWUE resulted in an increase in radial growth. These results suggest that spruce forests in southwestern China have not shown declining trends under increasing Ca and climate change scenarios, in contrast to trees growing in water-limited areas. Therefore, spruce forests benefit from the increasing CO₂ in the atmosphere in the humid areas of southwestern China.