In modeling forest stand growth and yield,crown width,a measure for stand density,is among the parameters that allows for estimating stand timber volumes.However,accurately measuring tree crown size in the field,in particu-lar for mature trees,is challenging.This study demonstrated a novel method of applying machine learning algorithms to aerial imagery acquired by an unmanned aerial vehi-cle (UAV) to identify tree crowns and their widths in two loblolly pine plantations in eastern Texas,USA.An ortho mosaic image derived from UAV-captured aerial photos was acquired for each plantation (a young stand before canopy closure,a mature stand with a closed canopy).For each site,the images were split into two subsets:one for training and one for validation purposes.Three widely used object detection methods in deep learning,the Faster region-based convolutional neural network (Faster R-CNN),You Only Look Once version 3 (YOLOv3),and single shot detection(SSD),were applied to the training data,respectively.Each was used to train the model for performing crown recogni-tion and crown extraction.Each model output was evaluated using an independent test data set.All three models were successful in detecting tree crowns with an accuracy greater than 93％,except the Faster R-CNN model that failed on the mature site.On the young site,the SSD model performed the best for crown extraction with a coefficient of determination(R2) of 0.92,followed by Faster R-CNN (0.88) and YOLOv3(0.62).As to the mature site,the SSD model achieved a R2 as high as 0.94,follow by YOLOv3 (0.69).These deep leaning algorithms,in particular the SSD model,proved to be successfully in identifying tree crowns and estimat-ing crown widths with satisfactory accuracy.For the pur-pose of forest inventory on loblolly pine plantations,using UAV-captured imagery paired with the SSD object deten-tion application is a cost-effective alternative to traditional ground measurement.

School of Information Engineering,Zhejiang A & F University,Lin'an 311300,Zhejiang,People's Republic of China;Key Laboratory of State Forestry and Grassland Administration On Forestry Sensing Technology and Intelligent Equipment,Zhejiang A & F University,Lin'an 311300,Zhejiang,People's Republic of China;Key Laboratory of Forestry Intelligent Monitoring and Information Technology of Zhejiang Province,Zhejiang A & F University,Lin'an 311300,Zhejiang,People's Republic of China;Jiyang College of Zhejiang A & F University,Zhuji 311800,Zhejiang,People's Republic of China;College of Forestry and Biotechnology,Zhejiang A & F University,Lin'an 311300,Zhejiang,People's Republic of China;College of Forestry and Agriculture,Stephen F.Austin State University,Nacogdoches,TX 75962,USA

[1]Xiongwei Lou;Yanxiao Huang;Luming Fang;Siqi Huang;Haili Gao;Laibang Yang;Yuhui Weng;I.-K.uai Hung-.Measuring loblolly pine crowns with drone imagery through deep learning)[J].林业研究（英文版）,2022(01):227-238

loblolly,estimat,deten

Measuring,pine,crowns,drone,imagery,through,deep,learning,In,modeling,forest,stand,growth,yield,density,among,parameters,that,allows,estimating,timber,volumes,However,accurately,measuring,size,field,mature,trees,challenging,This,study,demonstrated,novel,applying,machine,algorithms,aerial,acquired,by,unmanned,vehi,cle,UAV,their,widths,plantations,eastern,Texas,USA,An,ortho,mosaic,derived,from,captured,photos,was,each,young,before,canopy,closure,closed,For,site,images,were,split,into,subsets,training,validation,purposes,Three,widely,used,object,detection,methods,Faster,region,convolutional,neural,network,You,Only,Look,Once,version,YOLOv3,single,shot,SSD,applied,data,respectively,Each,performing,recogni,extraction,output,evaluated,using,independent,test,All,three,models,detecting,accuracy,greater,than,except,failed,performed,best,coefficient,determination,followed,achieved,high,These,leaning,particular,proved,successfully,identifying,satisfactory,inventory,paired,application,cost,effective,alternative,traditional,ground,measurement

0.484054