基于多特征融合和知识蒸馏的亚热带常见乔木识别方法  

Common arbor identification method in subtropics based on multiple features fusion and knowledge distillation

作　　者：丁鋆 徐爱俊[1,2,3] 吴小芬 周素茵[1,2,3] Ding Yun;Xu Aijun;Wu Xiaofen;Zhou Suyin(College of Mathematics and Computer Science,Zhejiang A&F University,Hangzhou 311300,China;Zhejiang Provincial Key Laboratory of Forestry Intelligent Monitoring and Information Technology,Zhejiang A&F University,Hangzhou 311300,China;Key Laboratory of National Forestry and Grassland Administration on Forestry Sensing Technology and Intelligent Engineering,Zhejiang A&F University,Hangzhou 311300,China;Hangzhou Lin′an District Rural Water Asset Management Co.,Ltd.,Hangzhou 311300,China)

机构地区：[1]浙江农林大学数学与计算机科学学院,浙江杭州311300 [2]浙江农林大学浙江省林业智能监测与信息技术研究重点实验室,浙江杭州311300 [3]浙江农林大学林业感知技术与智能装备国家林业与草原局重点实验室,浙江杭州311300 [4]杭州市临安区农村水务资产经营有限公司,浙江杭州311300

出　　处：《电子技术应用》2024年第8期1-9,共9页Application of Electronic Technique

基　　金：国家自然科学基金(32371867)。

摘　　要：乔木在维持生态平衡、保护生物多样性以及调节气候和改善空气质量等方面发挥着至关重要的作用。针对复杂背景下乔木识别准确率较低的问题,提出了一种基于树木多特征融合和知识蒸馏的亚热带常见乔木识别模型MFFMN-KD-TA。该模型采用3个并行的MobileNetV3_Small主干网络分别提取树叶、树干和树木整体特征;并通过知识蒸馏和嵌入Triplet Attention模块的方法优化训练。试验结果表明,MFFMN-KD-TA模型在自建树木测试集上的准确率、精确率和F1分数分别为0.9609、0.9621和0.9608,较MFFMN模型分别提升了3.05%、2.83%和3.07%。与三分支融合模型3-ShuffleNetV2和3-MobileNetV2相比,提出的多特征融合模型MFFMN-KD-TA参数量较小且能够较准确地识别乔木种类,为亚热带和其他地区的树种识别提供了新思路和新方法。Trees play a vital role in maintaining ecological balance,protecting biodiversity,regulating climate and improving air quality.In order to solve the problem of low tree identification accuracy in complex backgrounds,a tree species identification model MFFMN-KD-TA for common arbor in subtropics is proposed based on tree multi-feature fusion and knowledge distillation.The model uses three parallel MobileNetV3_Small backbone networks to extract features of leaves,trunks and overall trees respectively,and optimizes training by using knowledge distillation and embedding Triplet Attention modules.The test results show that the accuracy,precision and F1 score of the MFFMN-KD-TA model on the self-built tree test set are 0.9609,0.9621 and 0.9608 respectively,which are 3.05%,2.83%and 3.07%higher than the MFFMN model respectively.Compared with the threebranch fusion models 3-ShuffleNetV2 and 3-MobileNetV2,the multi-feature fusion model MFFMN-KD-TA proposed in this study has a smaller number of parameters and can identify arbor species more accurately,providing a new idea and method for tree species identification in subtropics and other areas.

关 键 词：树种识别 亚热带地区 MobileNetV3 多特征融合 知识蒸馏 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]