Tree Species Classification of Power Line Corridor Based on Multi-source Remote Sensing Data
-
摘要: 针对目前树冠提取中受背景影响和易出现过度分割的问题,首先,采用可见光差异植被指数和双边滤波对传统的单木树冠分割方法进行了改进;然后,以单木树冠为对象提取多维特征,并利用XGBoost算法进行特征重要性排序和特征选择;最后,使用随机森林、支持向量机、人工神经网络3种非参数分类器,设计了12种分类方案,进行了单木树种分类和精度评价。结果表明,改进的单木分割方法可以有效提高树冠提取精度,得到的树冠分割精度在80%以上;将Li DAR数据和航空正射影像相结合,采用XGBoost算法进行特征选择后,使用ANN分类器的分类方案精度最高,总体精度为86.19%,说明多源数据协同作用和特征选择可以提高树种分类精度,在单木尺度上ANN分类器对现有树种类型的分类能力最强。Abstract: The effectiveness of airborne Li DAR point cloud and aerial imagery on tree species classification and the effect of XGBoost algorithm for feature selection on tree species classification accuracy were researched,and the ability three non-parametric classifiers of random forest,support vector machine and artificial neural network to classify tree species on a single-wood scale were evaluated.Aiming at the current background effect of canopy extraction and the problem of over-segmentation,the traditional single tree canopy segmentation method was improved by using the visible light difference vegetation index and bilateral filtering; and then the single tree canopy was used as an object to extract multi-dimensional features by using the XGBoost algorithm to perform feature importance ranking and feature selection. Finally,three non-parameter classifiers of random forest,support vector machine and artificial neural network were used to design 12 classification schemes to classify single tree species and do accuracy evaluation. The results showed that the improved single tree segmentation method can effectively improve the accuracy of tree crown extraction,and the accuracy of the obtained tree canopy segmentation results was more than 80%; the Li DAR data and aerial orthophotos were combined,and the ANN classifier was used for feature selection after XGBoost algorithm for feature selection. The scheme had the highest accuracy,with an overall accuracy of 86. 19%,indicating that multi-source data synergy and feature selection can improve the accuracy of tree species classification. The ANN classifier had the strongest ability to classify existing tree species on a single tree scale.
-
-
[1] 穆超.基于多种遥感数据的电力线走廊特征物提取方法研究[D].武汉:武汉大学,2010.MU Chao. Research on extraction method of power line corridor features based on multiple remote sensing data[D]. Wuhan:Wuhan University,2010.(in Chinese) [2] 张大力.基于多光谱CCD影像和Li DAR数据的单木树种分类研究[D].哈尔滨:东北林业大学,2019.ZHANG Dali. Research on classification of single tree species based on multispectral CCD image and Li DAR data[D]. Harbin:Northeast Forestry University,2019.(in Chinese) [3] ZHANG Z,KAZAKOVA A,MOSKAL L,et al. Object-based tree species classification in urban ecosystems using Li DAR and hyperspectral data[J]. Forests,2016,7(12):122.
[4] 毛学刚,杜子涵,刘家倩,等.基于无人机Li DAR的天然林与人工林林隙提取[J/OL].农业机械学报,2020,51(3):232-240.MAO Xuegang,DU Zihan,LIU Jiaqian,et al. Extraction of forest gaps in natural forest and man-made forest based on UAV Li DAR[J/OL]. Transactions of the Chinese Society for Agricultural Machinery,2020,51(3):232-240. http:∥www. j-csam. org/jcsam/ch/reader/view_abstract. aspx? file_no=20200327&flag=1. DOI:10. 6041/j. issn. 1000-1298. 2020. 03. 027.(in Chinese) [5] 曹林,佘光辉.基于机载小光斑全波形Li DAR的亚热带林分特征反演[J].林业科学,2015,51(6):81-92.CAO Lin,SHE Guanghui. Inversion of forest stand characteristics using small-footprint[J]. Scientia Silvae Sinicae,2015,51(6):81-92.(in Chinese) [6] WANG K,WANG T,LIU X. A review:individual tree species classification using integrated airborne Li DAR and optical imagery with a focus on the urban environment[J]. Forests,2019,10(1):1-18.
[7] 徐凡,张雪红,石玉立.基于激光雷达和航拍影像的城市地物分类研究[J].遥感技术与应用,2019,34(2):253-262.XU Fan,ZHANG Xuehong,SHI Yuli. Research on classificion of land cover based on Li DAR cloud and arial images[J].Remote Sensing Technology and Application,2019,34(2):253-262.(in Chinese) [8] YANG J,JONES T,CASPERSEN J,et al. Object-based canopy gap segmentation and classification:quantifying the pros and cons of integrating optical and Li DAR data[J]. Remote Sensing,2015,7(12):15917-15932.
[9] LINDBERG E,HOLMGREN J. Individual tree crown methods for 3D data from remote sensing[J]. Current Forestry Reports,2017,3(1):19-31.
[10] 李峰.机载Li DAR点云数据分类方法研究[D].湘潭:湘潭大学,2018.LI Feng. Research on classification method of airborne Li DAR point cloud data[D]. Xiangtan:Xiangtan University,2018.(in Chinese) [11] 王平.基于机载Li DAR数据和航空像片的单木参数提取研究[D].哈尔滨:东北林业大学,2012.WANG Ping. Research on single wood parameter extraction based on airborne Li DAR data and aerial photos[D]. Harbin:Northeast Forestry University,2012.(in Chinese) [12] 王彬.基于机载Li DAR和高光谱数据的单木提取和树种识别[D].南京:南京信息工程大学,2018.WANG Bin. Single tree extraction and tree species recognition based on airborne Li DAR and hyperspectral data[D]. Nanjing:Nanjing University of Information Science and Technology,2018.(in Chinese) [13] 王欣,陈传法.Li DAR森林冠层高度模型凹坑去除方法[J].测绘科学,2016,41(12):157-161.WANG Xin,CHEN Chuanfa. Method for removing pits of canopy height model from airborne Li DAR data[J]. Science of Surveying and Mapping,2016,41(12):157-161.(in Chinese) [14] WU X,SHEN X,CAO L,et al. Assessment of individual tree detection and canopy cover estimation using unmanned aerial vehicle based light detection and ranging(UAV-Li DAR)data in planted forests[J]. Remote Sensing,2019,11(8):908-929.
[15] FASSNACHT F E,LATIF H,STERENCZAK K,et al. Review of studies on tree species classification from remotely sensed data[J]. Remote Sensing of Environment,2016,186:64-87.
[16] 于旭宅,王瑞瑞,陈伟杰.改进分水岭算法在无人机遥感影像树冠分割中的应用[J].福建农林大学学报(自然科学版),2018,47(4):428-434.YU Xuzhai,WANG Ruirui,CHEN Weijie. Forest canopy segmentation of UAV remote sensing images using improved watershed algorithm[J]. Journal of Fujian Agriculture and Forestry University(Natural Science Edition),2018,47(4):428-434.(in Chinese) [17] 汪小钦,王苗苗,王绍强,等.基于可见光波段无人机遥感的植被信息提取[J].农业工程学报,2015,31(5):152-159.WANG Xiaoqin,WANG Miaomiao,WANG Shaoqiang,et al. Extraction of vegetation information from visible unmanned aerial vehicle images[J]. Transactions of the CSAE,2015,31(5):152-159.(in Chinese) [18] 李明东,李雪竹,胡昊东,等.基于双边滤波的图像阈值降噪算法改进与研究[J].九江学院学报(自然科学版),2019(2):65-67.LI Mingdong,LI Xuezhu,HU Haodong,et al. Improvement and research on image threshold denoising algorithm based on bilateral filtering[J]. Journal of Jiujiang University(Natural Science Edition),2019(2):65-67.(in Chinese) [19] 朱双志.面向对象的高分辨率遥感图像分割方法的研究[D].长沙:湖南大学,2012.ZHU Shuangzhi. Research on object-oriented high-resolution remote sensing image segmentation method[D]. Changsha:Hunan University,2012.(in Chinese) [20] 张腾飞.基于e Cognition分类的森林蓄积量估测研究[D].西安:西安科技大学,2012.ZHANG Tengfei. Research on estimation of forest volume based on e Cognition classification[D]. Xi’an:Xi’an University of Science and Technology,2012.(in Chinese) [21] 宋宜昊.基于易康软件平台下的北京城区林木树冠覆盖解译与检验[D].北京:中国林业科学研究院,2016.SONG Yihao. Interpretation and inspection of forest tree canopy cover in Beijing urban area based on Yikang software platform[D]. Beijing:Chinese Academy of Forestry,2016.(in Chinese) [22] 郑鑫,王瑞瑞,靳茗茗.基于形态学阈值标记分水岭算法的高分辨率影像单木树冠提取[J].中南林业调查规划,2017,36(4):30-35.ZHENG Xin,WANG Ruirui,JIN Mingming. Extraction of high-resolution images of single tree crown based on watershed algorithm with morphological threshold mark[J]. Central South Forest Inventory and Planning,2017,36(4):30-35.(in Chinese) [23] 季金胜.高分辨率遥感影像典型地物目标的特征选择及其稳定性研究[D].上海:上海交通大学,2015.JI Jinsheng. Research on feature selection and stability of typical objects in high-resolution remote sensing images[D].Shanghai:Shanghai Jiao Tong University,2015.(in Chinese) [24] 李占山,刘兆赓.基于XGBoost的特征选择算法[J].通信学报,2019,40(10):101-108.LI Zhanshan,LIU Zhaogeng. Feature selection algorithm based on XGBoost[J]. Journal on Communications,2019,40(10):101-108.(in Chinese) [25] IMMITZER M,ATZBERGER C,KOUKAL T. Tree species classification with random forest using very high spatial resolution8-Band World View-2 satellite data[J]. Remote Sensing,2012,4(9):2661-2693.
[26] RACZKO E,ZAGAJEWSKI B. Comparison of support vector machine,random forest and neural network classifiers for tree species classification on airborne hyperspectral APEX images[J]. European Journal of Remote Sensing,2017,50(1):144-154.
[27] 路威.面向目标探测的高光谱影像特征提取与分类技术研究[D].郑州:中国人民解放军信息工程大学,2005.LU Wei. Research on feature extraction and classification technology of hyperspectral image for target detection[D].Zhengzhou:Chinese People’s Liberation Army Information Engineering University,2005.(in Chinese) [28] 任建斌.基于小波变换和BP人工神经网络的遥感影像分类研究[D].呼和浩特:内蒙古师范大学,2012.REN Jianbin. Research on remote sensing image classification based on wavelet transform and BP artificial neural network[D].Huhhot:Inner Mongolia Normal University,2012.(in Chinese)
计量
- 文章访问数: 10
- HTML全文浏览量: 0
- PDF下载量: 7
下载:
