数字农科院2.0

A semi-supervised framework for UAV-based individual tree crown segmentation in structurally heterogeneous planted forests

文献类型： 外文期刊

作者： Qing Wang;Yihui Zhao;Yingpu Che;Han Shen;Yongbin Qiu;Yixiang Wang

作者机构：

关键词： Individual tree crown segmentation;Remote sensing;Semi-supervised learning;Structural parameters;Structurally heterogeneous forests

期刊名称： International Journal of Applied Earth Observation and Geoinformation

ISSN： 1569-8432

年卷期： 2026 年 146 卷

页码：

收录情况： SCIE（2025版）

摘要： Accurate individual tree crown (ITC) segmentation is essential for quantifying forest fine carbon stocks. However, deep learning segmentation methods face prohibitive annotation costs while unsupervised algorithms struggle in structurally complex forests. This study proposes a semi-supervised framework integrating unsupervised pseudo-label generation, deep instance segmentation, and staged fine-tuning to overcome these limitations. The framework transforms outputs from marker-controlled watershed (MCWS) and region-growing (itcSegment) algorithms into initial training targets to obtain generalizable instance segmentation representations, then refines Mask R-CNN models with minimal manual annotations to enhance structural awareness of instance boundaries and semantic context understanding, significantly reducing labeling dependency while improving robustness. Validated across three structurally heterogeneous Chinese fir stands spanning age (11–67 years), density (450–2500 stems·ha−1), and elevation (192–1047 m) gradients, our UAV RGB-based framework achieved consistent superiority over LiDAR and fused inputs, with spectral-textural features proving dominant for boundary delineation. It attained F1-scores of 0.826 (young, undulating terrain), 0.836 (mature, high-density), and 0.711 (over-mature, occluded) using only 40 % expert annotations (0.6 personnel-hours), representing a 0.42 average improvement over MCWS/itcSegment baselines. The designed staged fine-tuning strategy effectively mitigated pseudo-label error propagation, while expanding this annotation effort to 70–100 % yielded marginal accuracy gains, demonstrating exceptional efficiency in leveraging minimal supervision. Based on the segmented crowns, structural parameters were extracted with high fidelity: crown diameter (R2 ≥ 0.76, rRMSE ≤ 11.3 %) and crown area (R2 ≥ 0.88, rRMSE ≤ 16.4 %). This approach reduces annotation demands while maintaining robustness across forest heterogeneity, enabling operationally scalable solutions for precision forestry.

分类号：

  • 相关文献

[1]Quantitative structure-toxicity relationship study of lethal concentration to tadpole (Bufo vulgaris formosus) for organophosphorous pesticides. Yan, Dongyun,Xu, Shaohui,Yan, Dongyun,Jiang, Xin,Bian, Yongrong,Yu, Guifen,Wang, Ligang.

[2]Semi-supervised learning and attention mechanism for weed detection in wheat. Teng Liu,Xiaojun Jin,Luyao Zhang,Jie Wang,Yong Chen,Chengsong Hu,Jialin Yu. 2023

[3]Self-paced semi-supervised feature selection with application to multi-modal Alzheimer's disease classification. Chao Zhang,Wentao Fan,Bo Wang,Chunlin Chen,Huaxiong Li. 2024

[4]Optimization and Test of Structural Parameters of Flat Hob Chopper. Anming Yang,Wei Xiang,Bo Yan,Yiping Duan,Jiangnan Lv,Mingliang Wu. 2022

[5]Transfer Learning in Junction With a Light Use Efficiency Model for Estimating Grassland Gross Primary Production. Yu, Ruiyang,Yao, Yunjun,Tang, Qingxin,Zhang, Xueyi,Shao, Changliang,Fisher, Joshua B.,Chen, Jiquan,Zhang, Xiaotong,Li, Yufu,Xu, Jia,Liu, Lu,Xie, Zijing,Ning, Jing,Fan, Jiahui,Zhang, Luna. 2025

[6]DEVELOPING A DESERTIFICATION ASSESSMENT SYSTEM USING A PHOTOSYNTHESIS MODEL WITH ASIMILLATED MULTI SATELLITE DATA. Kaneko, D.,Yang, P.,Kumakura, T.,Yang, P.,Chang, N. B.. 2010

[7]The Effect of Vegetation on the Remotely Sensed Soil Thermal Inertia and a Two-Source Normalized Soil Thermal Inertia Model for Vegetated Surfaces. Zhang, Renhua,Tian, Jing,Mi, Sujuan,Su, Hongbo,Liu, Kai,Mi, Sujuan,Liu, Kai,Su, Hongbo,He, Honglin,Li, Zhaoliang. 2016

[8]Detection of rice sheath blight for in-season disease management using multispectral remote sensing. Qin, ZH,Zhang, MH. 2005

[9]Spatial dynamics modelling of crops pattern with remote sensing classification data. Xia, Tian,Wu, Wenbin,Zhou, Qingbo,Yang, Peng,Liu, Yanxia. 2014

[10]Estimation of Atmospheric Profiles From Hyperspectral Infrared IASI Sensor. Wu, Hua,Tang, Bo-Hui,Ni, Li,Wang, Ning,Qian, Yonggang,Li, Zhao-Liang,Li, Zhao-Liang. 2013

[11]Monitoring the Plant Density of Cotton with Remotely Sensed Data. Bai, Junhua,Li, Shaokun,Bai, Junhua,Li, Jing,Bai, Junhua. 2011

[12]Assimilating MODIS-LAI into Crop Growth Model with EnKF to Predict Regional Crop Yield. Huang, Jianxi,Ma, Guannan,Wu, Sijie,Liu, Xingquan,Fan, Jinlong,Zou, Jinqiu. 2012

[13]Remote sensing monitoring of grass growing in grassland ecosystems of China. Qin, Zhihao,Xu, Bin,Tao, Weiguo. 2007

[14]Remote sensing of agro-droughts in Guangdong province of China using MODIS satellite data. Gao, Maofang,Qin, Zhihao,Yang, Xiuchun,Qin, Zhihao,Zhang, Hong'ou,Zhou, Xia,Qin, Zhihao,Lu, Liping. 2008

[15]Jointly Assimilating MODIS LAI and ET Products Into the SWAP Model for Winter Wheat Yield Estimation. Huang, Jianxi,Ma, Hongyuan,Su, Wei,Zhang, Xiaodong,Huang, Jianxi,Ma, Hongyuan,Su, Wei,Zhang, Xiaodong,Huang, Yanbo,Fan, Jinlong,Wu, Wenbin. 2015

[16]Progress in soil moisture estimation from remote sensing data for agricultural drought monitoring. Yan, Feng,Qin, Zhihao,Yan, Feng,Qin, Zhihao,Li, Maosong,Li, Wenjuan. 2006

[17]Monitoring land use dynamics for ecological degradation assessment in the rim zone of North China using MODIS and Land TM data. Qin, ZH,Xu, B,Liu, J,Li, WJ,Zhang, WC,Zhang, HO. 2004

[18]Estimation of Above-ground Biomass Carbon Storage in Hulunbeier Grassland based on Remotely Sensed Data. Hasituya,Chen Zhong-xin,Wu Wen-bin,Huang, Qing,Hasituya,Chen Zhong-xin,Wu Wen-bin,Huang, Qing. 2015

[19]Application of Crop Model Data Assimilation With a Particle Filter for Estimating Regional Winter Wheat Yields. Jiang, Zhiwei,Chen, Jin,Jiang, Zhiwei,Chen, Zhongxin,Liu, Jia,Ren, Jianqiang,Li, Zongnan,Sun, Liang,Li, He. 2014

[20]Spatiotemporal Variations in Grassland Desertification Based on Landsat Images and Spectral Mixture Analysis in Yanchi County of Ningxia, China. Li, Jinya,Li, Jinya,Xu, Bin,Zhao, Lina,Yang, Xiuchun,Jin, Yunxiang,Gao, Tian,Yu, Haida,Zhao, Fen,Ma, Hailong,Qin, Zhihao. 2014

作者其他论文 更多>>