基于UNet-ResNet14^(*)半监督学习的无人机影像森林树种分类  

Classification of tree species based on UNet-ResNet14^(*)semi-supervised learning using UAV images

作　　者：陈龙伟 周小成[1] 李传昕 林华章 王永荣 崔永红 CHEN Longwei;ZHOU Xiaocheng;LI Chuanxin;LIN Huazhang;WANG Yongrong;CUI Yonghong(Key Laboratory of Spatial Data Mining and Information Sharing at Fuzhou University,Ministry of Education,Fuzhou 350108,China;Fujian Datian Taoyuan State-owned Forest Farm,Sanming 366199,China;Fujian Zhangping Wuyi State-owned Forest Farm,Longyan 364400,China;Fujian Forest Inventory and Planning Institute,Fuzhou 350003,China)

机构地区：[1]福州大学空间数据挖掘与信息共享教育部重点实验室,福州350108 [2]福建省大田桃源国有林场,三明366199 [3]福建省漳平五一国有林场,龙岩364400 [4]福建省林业调查规划院,福州350003

出　　处：《农业工程学报》2024年第1期217-226,共10页Transactions of the Chinese Society of Agricultural Engineering

基　　金：福建省科技厅高校产学合作项目(2022N5008);福建省科技厅对外合作项目(2022I0007)。

摘　　要：无人机遥感在森林树种精细和高效分类制图中具有巨大的潜力。为了快速准确获取森林的优势树种分布信息,该研究探讨了半监督学习方法在树种分类方面的有效性。以福建省福州市、龙岩市和三明市的4个试验区为例,构建精简的ResNet18为主干的UNet树种分类模型(UNet-ResNet14^(*)),使用交叉熵和Dice系数的联合损失函数来优化模型参数,对比分析Self-training和Mean Teacher两种不同的半监督学习方法在无人机影像森林树种分类模型的泛化能力。结果表明,以ResNet14^(*)作为主干的分类模型与其他模型相比精度更高且预测速度更快,当联合损失函数权重值为0.5的情况下模型预测效果最好,总体精度达到了91.15%。经过Self-training的模型在木荷、马尾松、杉木3个样本充足的类别中精度均有所提升,总精度为91.08%,比原始模型略低,但在独立验证区的精度为88.50%,比原始模型高;Mean Teacher方法的总精度为88.56%,在独立验证区的精度为73.56%。因此,研究认为可以采用Self-trainin半监督方法结合UNet-ResNet14^(*)的方案快速得到试验区的树种组成信息。Unmanned aerial vehicle(UAV)remote sensing has the promising potential for the precise and efficient classification and mapping of forest tree species.Deep learning also requires a large number of datasets for training,typically on manual annotation.In this study,the framework of forest tree species classification was proposed to fully utilize a large amount of unlabeled data and a small amount of annotated data using semi-supervised learning.A rapid and accurate classification was also achieved in the high-precision distribution of dominant tree species in forests.The experimental areas were taken as the complex mountainous forest environment in Fujian Province.The composition of tree species was then obtained in a rapid,effective,and cost-saving manner.Taking four experimental areas in Fuzhou,Longyan,and Sanming in Fujian Province as examples,the simplified classification was constructed in the UNet tree species(ResNet14^(*))model with ResNet18 as the backbone.ResNet14^(*)was different from ResNet18:ResNet14^(*)was used to remove the layer4 part of ResNet18,i.e.,the last downsampled cascaded block,which retained slightly higher spatial information;At the end of the layer2 and layer3 sections of ResNet14^(*),a max pooling layer was added to reduce the training parameters of the neural network while retaining the original features.A joint loss function of cross entropy and Dice coefficient was used to optimize the model parameters.The generalization of Self-training and Mean teacher was evaluated on the classification models with semi-supervised learning using UVA images.The results show that the overall accuracy of the ResNet14^(*)network reached 91.15%,with a Kappa coefficient of 0.827,which was within 1%of the accuracy of the rest ResNet models.At the same time,a smaller number of parameters and the shortest prediction time were achieved to balance the accuracy and efficiency of tree species classification.The best prediction performance of ResNet14^(*)was achieved with the joint loss function weight of 0.5,in

关 键 词：无人机 遥感 森林 树种分类 可见光 UNet ResNet 半监督学习 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置] S757.2[自动化与计算机技术—控制科学与工程]