MDC-Net:a multi-directional constrained and prior assisted neural network for wood and leaf separation from terrestrial laser scanning

作     者：Wenxia Dai Yiheng Jiang Wen Zeng Ruibo Chen Yongyang Xu Ningning Zhu Wen Xiao Zhen Dong Qingfeng Guan 

作者机构：School of Geography and Information EngineeringChina University of GeosciencesWuhanPeople’s Republic of China Guangxi Zhuang Autonomous Region Institute of Natural Resources Remote SensingNanningPeople’s Republic of China School of Computer ScienceChina University of GeosciencesWuhanPeople’s Republic of China State Key Laboratory of Information Engineering in SurveyingMapping and Remote SensingWuhan UniversityWuhanPeople’s Republic of China National Engineering Research Center of Geographic Information SystemChina University of GeosciencesWuhanPeople’s Republic of China 

出 版 物：《International Journal of Digital Earth》 (国际数字地球学报（英文）)

年 卷 期：2023年第16卷第1期

页      面：1224-1245页

核心收录：

学科分类：12[管理学] 0907[农学-林学] 08[工学] 0710[理学-生物学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 0903[农学-农业资源与环境] 0804[工学-仪器科学与技术] 0901[农学-作物学] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0713[理学-生态学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：supported by the National Natural Science Foundation of China[grant number 42101456] funded by Key Laboratory of Spatial-temporal Big Data Analysis and Application of Natural Resources in Megacities,MNR(No.KFKT-2022-04) Open Research Fund of State Key Laboratory of Information Engineering in Surveying,Mapping and Remote Sensing of Wuhan University(21S01) Research Fund of post-doctoral innovation in Hubei Province under Grant No.1232168 

主　　题：Terrestrial laser scanning wood and leaf separation deep learning prior features 

摘      要：Wood-leaf separation from terrestrial laser scanning(TLS)is a crucial prerequisite for quantifying many biophysical properties and understanding ecological *** this study,we propose a novel multi-directional collaborative convolutional neural network(MDC-Net)that takes the original 3D coordinates and useful features from prior knowledge(prior features)as input,and outputs the semantic labels of TLS point *** MDC-Net contains two key units:(1)a multi-directional neighborhood construction(MDNC)unit to obtain more representative neighbors and enable directionally aware feature encoding in the subsequent local feature extraction,to mitigate occlusion effects;(2)a collaborative feature encoding(CFE)unit is introduced to incorporate useful features from prior knowledge into the network through a collaborative cross coding to enhance the discrimination for thin structures(*** branches and leaf).The MDC-Net is evaluated onfive plots from forests in Guangxi,China,with different branch architectures and leaf *** results showed that the MDC-Net achieved an OA of 0.973 and a mIoU of 0.821 and outperformed other related *** believe the MDC-Net would facilitate the usage of TLS in ecology studies for quantifying tree size and morphology and thus promote the development of relevant ecological applications.