摘要
为快速准确地监测红树林群落类型和分布,获得无人机影像在高精度红树林群落制图中的应用方法,使用无人机获取海南省澄迈县富力湾红树林国家级湿地公园的高空间分辨率(5 cm)可见光影像,应用面向对象分类和最优分割算法提取红树林冠层边界,使用可见光波段差异植被指数和随机森林分类算法识别红树林冠层群落。研究结果表明:面向对象的随机森林红树林群落分类中,选用最优分割尺度和适宜的随机森林算法特征和参数能够提高红树林群落分类精度,最终分类结果总体精度达89.09%,Kappa系数为0.87。该方法适合高精度、小尺度的红树林群落制图,能够及时、准确地监测红树林群落变化。因此,无人机可见光影像在红树林的监测与管理中有广阔的应用前景。
China′s Fuliwan National Mangrove Park,which is located in Chengmai County,Hainan Province,was selected to study methods for the regular and accurate monitoring of mangrove community species and distribution:unmanned aerial vehicle(UAV)imagery with high-spatial resolution was applied to achieve precise mangrove community mapping.A UAV capable of providing high-spatial resolution(5 cm),multispectral images(red,green,and blue bands)was used in this study.First,an object-oriented classification method and an optimal segmentation scale model were used to segment the mangrove canopies.Thereafter,the visible-band difference vegetation index(VDVI)and random forest classifier were applied to identify different mangrove canopy species.The results demonstrated that using optimal segmentation scale with the most suitable parameters and features in the random forest model facilitated accurate mangrove species classification.The overall accuracy and Kappa index of the classification results were 89.09%and 0.87,respectively,suggesting that this method is useful for small-scale,high-precision mapping of mangrove species while regularly and precisely monitoring changes.According to the results and discussion of this study,visible unmanned aerial vehicle imagery has enormous potential for managing mangrove ecosystems.
作者
闻馨
贾明明
李晓燕
王宗明
钟才荣
冯尔辉
WEN Xin;JIA Mingming;LI Xiaoyan;WANG Zongming;ZHONG Cairong;FENG Erhui(College of Earth Science,Jilin University,Changchun,Jilin 130061,China;Key Laboratory of Wetland Ecology and Environment,Northeast Institute of Geography and Agroecology,Chinese Academy of Sciences,Changchun,Jilin 130102,China;National Earth System Science Data Center,Beijing 100101,China;Hainan Dongzhaigang National Nature Reserve Administration Bureau,Haikou,Hainan 571129,China)
出处
《森林与环境学报》
CSCD
北大核心
2020年第5期486-496,共11页
Journal of Forest and Environment
基金
国家科技部基础资源调查专项(2017FY100706)
吉林省自然科学基金项目(20200201048JC)。
关键词
无人机
可见光影像
红树林冠层群落
随机森林分类
最优分割尺度
unmanned aerial vehicle
visible imagery
mangrove canopy species
random forest classification
optimal segmentation scale
