基于GF-2遥感影像的澳洲坚果林空间分布信息提取  

Spatial distribution information extraction of macadamia forest based on GF-2 remote sensing image

作　　者：王耀磊 郑毅 张成程 荣渝虹 梁启斌[5] 王艳霞[1] 侯磊[5] 李晓琳[1] WANG Yao-lei;ZHENG Yi;ZHANG Cheng-cheng;RONG Yu-hong;LIANG Qi-bin;WANG Yan-xia;HOU Lei;LI Xiao-lin(College of Soil and Water Conservation,Southwest Forestry University,Kunming,Yunnan 650224,China;Yunnan Open University,Kunming,Yunnan 650500,China;Southwest Investigation and Planning Institute,National Forestry and Grassland Administration,Kunming,Yunnan 650021,China;Yunnan Institute of Tropical Crops,Jinghong,Yunnan 666100,China;College of Ecology and Environment,Southwest Forestry University,Kunming,Yunnan 650224,China)

机构地区：[1]西南林业大学水土保持学院,云南昆明650224 [2]云南开放大学,云南昆明650500 [3]国家林业和草原局西南调查规划院,云南昆明650021 [4]云南省热带作物科学研究所,云南景洪666100 [5]西南林业大学生态与环境学院,云南昆明650224

出　　处：《南方农业学报》2025年第1期74-86,共13页Journal of Southern Agriculture

基　　金：云南省农业基础研究联合专项(202101BD070001-111);云南省重大科技专项—林草科技创新联合专项(202404CB090001);云南省国有自然资源资产权益管理试点项目(632171);云南云天化股份有限公司项目(YTH-4320-WB-2021-037666-00)。

摘　　要：【目的】基于GF-2遥感影像快速准确获取澳洲坚果林的空间分布信息,为有效利用GF-2遥感影像研究西南山区澳洲坚果林分布及为山地丘陵区其他地物类型信息的提取提供参考依据。【方法】以云南省临沧市镇康县南伞镇为研究区,GF-2影像和数字高程模型(DEM)为数据源。通过面向对象的方法,提取影像对象的光谱特征、纹理特征、形状特征和地形特征共90维特征变量,设计8种组合方案(方案A1~方案A8),使用平均不纯度减少的方法对特征重要性进行度量,选取最佳特征组合,采用随机森林、支持向量机和决策树算法对澳洲坚果林进行提取,探讨不同特征类型和分类算法对澳洲坚果林提取精度的影响。【结果】相比遍历分割参数法,尺度参数估算(ESP)工具和邻域差分绝对值与标准差比(RMAS)结合的方法能够更高效、客观地确定特定地物的最佳分割尺度;通过对比方案A8和方案A7可知,方案A8中加入地形特征后,整体特征维度有所降低,主要表现为纹理特征数量减少,仅保留4个纹理特征。不同类型特征对澳洲坚果林识别的贡献排序为光谱特征>地形特征>纹理特征>形状特征。在分类算法角度方面,随机森林在总体精度(OA)、用户精度(UA)、生产者精度(PA)和Kappa系数等精度指标上均优于支持向量机和决策树,方案A8融合了所有特征取得最佳的分类效果,4个指标均高于其他方案。光谱特征、纹理特征、形状特征和地形特征组合的随机森林分类方法精度最佳,OA达95.8%,澳洲坚果林的PA为87.7%,UA为94.3%。澳洲坚果林空间分布特征结果显示,澳洲坚果在15°~20°坡度范围的种植面积最大,为2.9 km^(2);澳洲坚果林面积主要分布在东南坡向和900~1200 m的海拔范围内。【结论】地形+纹理+形状+地形组合方案经特征优选后结合随机森林算法,能够有效识别澳洲坚果林的分布。GF-2遥感数据与面向对象法在【Objective】This study aimed to quickly and accurately obtain the spatial distribution information of macadamia forests based on GF-2 remote sensing image.It provided reference for the effective utilization of GF-2 remote sensing image to study the distribution of macadamia forests in the southwestern mountainous areas,as well as for the extraction of other land cover types in hilly and mountainous regions.【Method】The study area was located in Nansan Town,Zhenkang County,Lincang City,Yunnan Province.GF-2 image and digital elevation model(DEM)were used as data sources.An object-oriented approach was employed to extract 90 dimensional feature variables,including spectral,texture,shape and terrain features.Eight feature combination schemes(A1 to A8)were designed.The importance of the features was measured using the mean decrease in impurity(MDI)method,and the best feature combination was selected.Random forest(RF),support vector machine(SVM),and decision tree(DT)algorithms were used for the extraction of macadamia nut forests.The study explored the influence of different feature types and classification algorithms on the accuracy of macadamia nut forest extraction.【Result】Compared to the exhaustive segmentation parameter method,the combination of the scale parameter estimation(ESP)tool and the neighborhood difference absolute value and standard deviation ratio(RMAS)method was more efficient and objective in determining the optimal segmentation scale for specific land cover types.By comparing scheme A8 with scheme A7,it was found that adding terrain as a feature in scheme A8 reduced the overall feature dimensionality,particularly in the texture features,with only 4 texture features retained.The contribution of different feature types to the macadamia nut forest identification was ranked as follows:spectral features>terrain features>texture features>shape features.In terms of classification algorithms,random forest outperformed support vector machine and decision tree in overall accuracy(OA),user accuracy(U

关 键 词：澳洲坚果 GF-2遥感影像 面向对象 特征优选 随机森林 

分 类 号：S664.939[农业科学—果树学]