文献类型：期刊文献

中文题名：基于资源三号的雷州湾红树林分类研究

英文题名：Classification of Mangrove in Leizhou Bay Based on ZY-3

作者：王文芳[1];董昭顷[1];付东洋[1];祁雅莉[1];林道荣[1]

机构：[1]广东海洋大学电子与信息工程学院,广东湛江524088

年份：2020

卷号：40

期号：1

起止页码：35

中文期刊名：海洋测绘

外文期刊名：Hydrographic Surveying and Charting

收录：CSTPCD、、CSCD_E2019_2020、CSCD

基金：国家海洋公益专项(201305019);广东省自然科学基金(2014A030313603);广东省科技计划(2013B030200002,2016A020222016);广东海洋大学创新强校项目(GDOU2014050226,GDOU2014050246);广东海洋大学大学生创新创业训练计划(CXXL2017026);“海之帆”起航计划大学生科技创新培育项目(qhjh2017zr15).

语种：中文

中文关键词：资源三号;红树林分类;面向对象;最优参数组合;雷州湾

外文关键词：ZY-3;classification of mangrove;object-oriented;optimal parameter combination;leizhou bay

中文摘要：基于面向对象的分类方法,不同参数组合会对红树林分类精度产生影响。以雷州半岛东岸附城镇沿海一带为研究区域,探索最优的参数组合以实现红树林的精确分类。利用资源三号(ZY-3)高分影像,基于图像光谱、形状和空间关系特征,对红树林进行分层次提取。结合红树林种类的光谱、空间特征差异,对比分析面向对象方法下不同因子、分割尺度及分类器对应下的分类精度,得出该研究区红树林树种在面向对象分类方法中的最优参数组合。结果表明:基于形状因子0.6+紧致度因子0.6、分割尺度为46的条件下,随机树分类器能有效区分无瓣海桑、白骨壤和秋茄三种红树林,总体精度为87.55%,Kappa系数为0.81。

外文摘要：Different parameter combinations will affect the accuracy of mangrove classification based on the object-oriented classification method.The eastern coastal Fucheng town of Leizhou Peninsula is taken as the research area to explore the optimal parameter combinations to achieve the accurate classification of mangrove.Mangroves are extracted hierarchically based on spectral,shape and spatial relationship characteristics of the high-resolution image of ZY-3.Combining the differences of spectral and spatial characteristics among mangrove species,the classification accuracy which of different factors,segmentation scales and classifiers under object-oriented method is compared and analyzed.The optimal parameter combination of the mangrove species in the study area is obtained in the object-oriented classification method.The experimental results show that based on the shape factor 0.6+compactness factor of 0.6 and the segmentation scale of 46,random tree classifier can effectively distinguish three mangroves,including Sonneratia apetala,Avicennia marina and Kandelia,with an overall precision of 87.55%and a Kappa coefficient of 0.81.