文献类型：期刊文献

英文题名：PosE-Enhanced Point Transformer with Local Surface Features (LSF) for Wood-Leaf Separation

作者：Lu, Xin[1,2] Wang, Ruisheng[3] Zhang, Huaiqing[4] Zhou, Ji[5] Yun, Ting[1,2]

第一作者：Lu, Xin

通信作者：Yun, T[1];Yun, T[2]

机构：[1]Nanjing Forestry Univ, Coinnovat Ctr Sustainable Forestry Southern China, Nanjing 210037, Peoples R China;[2]Nanjing Forestry Univ, Coll Informat Sci & Technol & Artificial Intellige, Nanjing 210037, Peoples R China;[3]Univ Calgary, Dept Geomatics Engn, Calgary, AB T2N 1N4, Canada;[4]Chinese Acad Forestry, Res Inst Forest Resources Informat Tech, Beijing 100091, Peoples R China;[5]Natl Inst Agr Bot NIAB, Cambridge Crop Res, Cambridge CB3, England

年份：2024

卷号：15

期号：12

外文期刊名：FORESTS

收录：;EI(收录号：20245317604494);Scopus(收录号：2-s2.0-85213280513);WOS:【SCI-EXPANDED(收录号:WOS:001384335100001)】；

基金：This study was financially supported by the National Natural Science Foundation of China under Grants 32371876, 32271877, and 42101451; the Natural Science Foundation of Jiangsu Province, China under Grant BK20221337; the Jiangsu Provincial Agricultural Science and Technology Independent Innovation Fund Project under Grant CX(22)3048; and the Key Laboratory of Land Satellite Remote Sensing Application, Ministry of Natural Resources of the People's Republic of China under Grant KLSMNR-G202208.

语种：英文

外文关键词：deep learning; wood-leaf separation; point clouds; forest digital twin; Local Surface Features; Point Feature Histogram

摘要：Wood-leaf separation from forest LiDAR point clouds is a challenging task due to the complex and irregular structures of tree canopies. Traditional machine vision and deep learning methods often struggle to accurately distinguish between fine branches and leaves. This challenge arises primarily from the lack of suitable features and the limitations of existing position encodings in capturing the unique and intricate characteristics of forest point clouds. In this work, we propose an innovative approach that integrates Local Surface Features (LSF) and a Position Encoding (PosE) module within the Point Transformer (PT) network to address these challenges. We began by preprocessing point clouds and applying a machine vision technique, supplemented by manual correction, to create wood-leaf-separated datasets of forest point clouds for training. Next, we introduced Point Feature Histogram (PFH) to construct LSF for each point network input, while utilizing Fast PFH (FPFH) to enhance computational efficiency. Subsequently, we designed a PosE module within PT, leveraging trigonometric dimensionality expansion and Random Fourier Feature-based Transformation (RFFT) for nuanced feature analysis. This design significantly enhances the representational richness and precision of forest point clouds. Afterward, the segmented branch point cloud was used to model tree skeletons automatically, while the leaves were incorporated to complete the digital twin. Our enhanced network, tested on three different types of forests, achieved up to 96.23% in accuracy and 91.51% in mean intersection over union (mIoU) in wood-leaf separation, outperforming the original PT by approximately 5%. This study not only expands the limits of forest point cloud research but also demonstrates significant improvements in the reconstruction results, particularly in capturing the intricate structures of twigs, which paves the way for more accurate forest resource surveys and advanced digital twin construction.