自适应集成的高分辨率树种遥感分类  

Adaptive Integrated High-resolution Tree Species Remote Sensing Classification

作　　者：蒲涛 王妮[4] 汤连盟 龚育红 王安 PU Tao;WANG Ni;TANG Lianmeng;GONG Yuhong;WANG An(School of Geomatics,Anhui University of Science and Technology,Huainan,Anhui 232001,China;Key Laboratory of Aviation-aerospace-ground Cooperative Monitoring and Early Warning of Coal Mining-induced Disasters of Anhui Higher Education Institutes,Huainan,Anhui 232001,China;Coal Industry Engineering Research Center of Mining Area Environmental and Disaster Cooperative Monitoring,Anhui University of Science and Technology,Huainan,Anhui 232001,China;School of Geographic Information and Tourism,Chuzhou University,Chuzhou,Anhui 239000,China;School of Resources and Environmental Engineering,Anhui University,Hefei 230601,China)

机构地区：[1]安徽理工大学空间信息与测绘工程学院,安徽淮南232001 [2]安徽理工大学矿山采动灾害空天地协同监测与预警安徽普通高校重点实验室,安徽淮南232001 [3]安徽理工大学矿区环境与灾害协同监测煤炭行业工程研究中心,安徽淮南232001 [4]滁州学院地理信息与旅游学院,安徽滁州239000 [5]安徽大学资源与环境工程学院,合肥230601

出　　处：《遥感信息》2023年第6期139-147,共9页Remote Sensing Information

基　　金：高校优秀青年人才支持计划一般项目(gxyq2021217);安徽省高校自然科学项目重点项目(KJ2021A1075);安徽理工大学研究生创新基金(2022CX2163)。

摘　　要：针对单一分类器无法均衡所有树种特征差异,基于指数函数、基分类器总精度(OA)及其变化趋势,提出一种用于集成多树种分类器的自适应权重分配方法(AWOL)。AWOL利用指数函数分布特点,以基分类器当前OA作为指数函数自变量分配基础权重,再引入OA计算的一阶导数作为约束因子放缩权重,使集成分类器进一步权衡树种特征分布,最后不断迭代训练自适应调整融合权重。结果表明,AWOL更能更有效地权衡不同树种特征,获得最佳树种分类结果,OA达到86.1%,训练与验证损失仅为0.3%。AWOL下的集成分类器与SLIC结合所绘制的树种分布图在所选人工林内制图OA最高达到了90%。分析证明,AWOL可进一步平衡树种特征差异,减缓过度拟合等现象,与分割算法结合可绘制更高精度树种分布图。This study aims to improve the accuracy of classifying tree species in a plantation forest.It develops a method called AWOL,which assigns weights to the base classifier based on its accuracy and uses first-order derivatives to adjust these weights.This allows the integrated classifier to better weigh the various features of different tree species and produce more accurate results.The study finds that AWOL can achieve an overall accuracy of 86.1%with a training and validation loss of just 0.3%.When combined with a segmentation method called SLIC,AWOL produces tree species distribution maps with a maximum accuracy of 90%.The results suggest that AWOL can better balance the variations in tree species features and prevent overfitting,resulting in more precise maps of tree species distribution.

关 键 词：AWOL 集成学习 深度学习 投票机制 树种分类 

分 类 号：P327[天文地球—地球物理学]