Article
Environmental Sciences
Jin Huang, Jantien Stoter, Ravi Peters, Liangliang Nan
Summary: This paper presents a fully automatic approach for reconstructing compact 3D building models from large-scale airborne point clouds. The approach addresses the challenge of missing vertical walls by inferring them directly from the data. The method outperforms state-of-the-art methods in terms of reconstruction accuracy and robustness, as demonstrated in experiments on various large-scale airborne LiDAR point clouds. Additionally, the authors have generated a new dataset with their method, which can stimulate research in urban reconstruction and the use of 3D city models in urban applications.
Article
Chemistry, Multidisciplinary
Young-Ha Shin, Kyung-Wahn Son, Dong-Cheon Lee
Summary: This paper focuses on utilizing the multiple returns in LiDAR data for building extraction using PointNet++. The experimental results show improved performance in building extraction. However, the method is limited by the lower point density in new data.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Software Engineering
Guillaume Coiffier, Justine Basselin, Nicolas Ray, Dmitry Sokolov
Summary: Surface reconstruction is crucial in processing point clouds, as it allows for inferring large-scale information. The template-based method with an energy function enables precise fitting and outlier segmentation in a single step.
COMPUTER-AIDED DESIGN
(2021)
Article
Environmental Sciences
Huxiong Li, Weiya Ye, Jun Liu, Weikai Tan, Saied Pirasteh, Sarah Narges Fatholahi, Jonathan Li
Summary: This study introduces a novel workflow for automated Digital Terrain Model (DTM) extraction from Airborne LiDAR point clouds based on a convolutional neural network and transfer learning. The results demonstrate that the proposed workflow establishes a superior DTM extraction accuracy with a root mean square error of only 7.3 cm for the interpolated DTM at 1 m resolution.
Article
Environmental Sciences
Pingbo Hu, Yiming Miao, Miaole Hou
Summary: This paper presents an automated modeling approach to semantically decompose and reconstruct complex building LiDAR point clouds into simple parametric structures, reducing the difficulty of interpreting complex building models. The proposed method is capable of efficiently decomposing complex building models into interpretable semantic structures.
Article
Remote Sensing
Chuan Zhao, Haitao Guo, Jun Lu, Donghang Yu, Xin Zhou, Yuzhun Lin
Summary: This paper proposes a new roof segmentation method for airborne LiDAR point clouds, which extracts reliable roof patches using a novel region growing strategy and RANSAC algorithm, and refines unsegmented points and segmentation results through iteration and voting, achieving accurate segmentation of roofs with different complexity and sizes.
REMOTE SENSING LETTERS
(2021)
Article
Environmental Sciences
Libo Cheng, Rui Hao, Zhibo Cheng, Taifeng Li, Tengxiao Wang, Wenlong Lu, Yulin Ding, Han Hu
Summary: This study focuses on the challenges of ground filtering in large scenes and introduces an elevation offset-attention (E-OA) module that considers global semantic features and integrates them into existing network frameworks. The experimental results demonstrate that this module significantly improves the ground filtering performance and surpasses traditional methods and other competing attention frameworks.
Article
Environmental Sciences
Marko Bizjak, Domen Mongus, Borut Zalik, Niko Lukac
Summary: This paper presents a novel automatic building reconstruction methodology based on half-spaces and height jump analysis. The methodology includes three stages: preprocessing, sub-building division, and reconstruction with half-spaces. The performance of the methodology was demonstrated on a large scale and validated on an ISPRS benchmark dataset.
Article
Remote Sensing
Huifang Feng, Yiping Chen, Zhipeng Luo, Wentao Sun, Wen Li, Jonathan Li
Summary: This paper proposes a preprocessing-free method for building instance extraction, which utilizes dual-channel airborne LiDAR data for point cloud processing. By reorganizing point cloud, rasterizing and constraint-based labeling, the method overcomes the challenges of new LiDAR data and improves extraction performance.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2022)
Article
Construction & Building Technology
Jie Shao, Wuming Zhang, Aojie Shen, Nicolas Mellado, Shangshu Cai, Lei Luo, Nan Wang, Guangjian Yan, Guoqing Zhou
Summary: The proposed method utilizes a novel top-down strategy to detect seed point sets with semantic information through cloth simulation, avoiding oversegmentation issues caused by a single seed point. Validation on three point cloud datasets shows high accuracy in roof extraction and effective reduction in error rates.
AUTOMATION IN CONSTRUCTION
(2021)
Article
Remote Sensing
Wangshan Yang, Xinyi Liu, Yongjun Zhang, Yi Wan, Zheng Ji
Summary: Building instance segmentation is crucial for parallel reconstruction, management, and analysis of building instances. Existing studies have mainly focused on building scenes with large building spacing, leading to low accuracy in complex building scenes and building point clouds. To address this, we propose a novel object-based building instance segmentation method using airborne LiDAR point clouds. The method divides point clouds into objects, classifies them based on roof plane, roof accessory, and building facade characteristics, and merges objects to obtain building instances.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2022)
Article
Forestry
Tauri Arumae, Mait Lang, Allan Sims, Diana Laarmann
Summary: The study aimed to predict the need for commercial thinning using airborne lidar data and the random forest machine learning algorithm. By testing forest stands in different regions of Estonia, it was found that the 95th height percentile and canopy cover were the most important metrics for predicting thinning needs. The prediction accuracy of this method was higher than that of the traditional linear model, with regional variations.
Article
Geography
Fayez Tarsha Kurdi, Mohammad Awrangjeb, Nosheen Munir
Summary: This article presents a new method for automatic building footprint modeling using only airborne LiDAR data, which consists of data filtering and roof modeling. By utilizing histogram interpretation and classification of roof plane boundaries, the accuracy of building footprint modeling has been significantly improved.
TRANSACTIONS IN GIS
(2021)
Article
Geography, Physical
Ruiqi Ma, Chi Chen, Bisheng Yang, Deren Li, Haiping Wang, Yangzi Cong, Zongtian Hu
Summary: In this paper, a corner-guided anchor-free single-stage 3D object detection model is proposed to estimate the 3D bounding boxes of objects by aggregating incomplete surface point clouds. The experiments demonstrate that this method achieves state-of-the-art performance in 3D object detection.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2022)
Article
Remote Sensing
Haoyi Xiu, Xin Liu, Weimin Wang, Kyoung-Sook Kim, Takayuki Shinohara, Qiong Chang, Masashi Matsuoka
Summary: Collapsed buildings need to be detected immediately after earthquakes for humanitarian assistance and post-disaster recovery. Automatic collapsed building detection using deep learning has gained popularity due to its superior ability to obtain discriminative feature representations. This study proposes a dedicated approach called Damage-Sensitive Network (DS-Net) for collapsed building detection by enhancing the feature representation of the damaged part using Laplacian Unit (LU) designed for 3D point clouds.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2023)
Article
Geochemistry & Geophysics
Yongjun Zhang, Xinyi Liu, Yi Zhang, Xiao Ling, Xu Huang
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2019)
Article
Geography, Physical
Yi Wan, Yongjun Zhang, Xinyi Liu
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2019)
Article
Environmental Sciences
Xinyi Liu, Yongjun Zhang, Xiao Ling, Yi Wan, Linyu Liu, Qian Li
Article
Geography, Physical
Dong Wei, Yongjun Zhang, Xinyi Liu, Chang Li, Zhuofan Li
Summary: A novel graph-based algorithm GLSM is proposed in this paper for line segment matching in two or multiple views, which can achieve sufficient matches and performs well on large image datasets. The implementation of GLSM will be available soon at the given website.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2021)
Article
Environmental Sciences
Yongjun Zhang, Wangshan Yang, Xinyi Liu, Yi Wan, Xianzhang Zhu, Yuhui Tan
Summary: This paper proposes a novel unsupervised building instance segmentation (UBIS) method that combines clustering algorithm and a model consistency evaluation method to address the issues of existing instance segmentation methods. Experimental results demonstrate that the proposed UBIS method performs well in various building scenes and outperforms state-of-the-art methods.
Article
Computer Science, Information Systems
Shuai Ling, Zhe Yu, Shaosheng Cao, Haipeng Zhang, Simon Hu
Summary: Transportation demand forecasting is crucial for optimal online transportation dispatch. However, the spatio-temporal complexity poses challenges, including compound spatial relationships, heterogeneity, and synchronicity between spatial and temporal relationships. To address these issues, we propose the STHAN framework, which explicitly captures compound spatial relationships using meta-paths. It constructs a spatio-temporal heterogeneous graph and utilizes hierarchical attention to capture heterogeneity and model the synchronicity between temporal relationships and spatial relationships.
ACM TRANSACTIONS ON KNOWLEDGE DISCOVERY FROM DATA
(2023)
Article
Geochemistry & Geophysics
Siyuan Zou, Xinyi Liu, Xu Huang, Yongjun Zhang, Senyuan Wang, Shuang Wu, Zhi Zheng, Bingxin Liu
Summary: This letter presents a LiDAR and image line-guided stereo matching method (L2GSM) that combines sparse but high-accuracy LiDAR points and sharp object edges of images to produce accurate and detailed point clouds. By using LiDAR depth information to extract depth discontinuity lines on the image, trilateral update of cost volume and depth discontinuity lines-aware semi-global matching (SGM) strategies are proposed to incorporate LiDAR data and depth discontinuity lines into the dense matching algorithm. Experimental results on indoor and aerial datasets demonstrate that our method greatly enhances the performance of the original SGM and surpasses two state-of-the-art LiDAR constraints' SGM methods, particularly in recovering the 3-D structure of low-textured and depth discontinuity regions. Moreover, the 3-D point clouds generated by our proposed method outperform the LiDAR data and dense matching point clouds generated by Metashape and SURE aerial in terms of completeness and edge accuracy.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2023)
Article
Geography, Physical
Xinyi Liu, Xianzhang Zhu, Yongjun Zhang, Senyuan Wang, Chen Jia
Summary: In this paper, a novel method for generating concise building models from dense meshes is proposed. The method involves extracting and completing planar primitives of the building, and reconstructing a concise polygonal mesh through connectivity-based primitive assembling. The method demonstrates high efficiency and robustness.
PHOTOGRAMMETRIC RECORD
(2023)
Article
Environmental Sciences
Xiaojian Liu, Yansheng Li, Xinyi Liu, Huimin Zou
Summary: This paper proposes a method for dark spot detection based on superpixels and deeper graph convolutional networks (SGDCNs), aiming to improve the accuracy of oil slick detection. The proposed method detects the contours of dark spots using superpixel segmentation and smooths the noise in SAR images. By selecting an appropriate subset of superpixel features using the support vector machine recursive feature elimination (SVM-RFE) algorithm, the learning task difficulty is reduced. Experimental results demonstrate the robustness and effectiveness of the proposed SGDCN method.
Article
Remote Sensing
Xiaojian Liu, Yongjun Zhang, Huimin Zou, Fei Wang, Xin Cheng, Wenpin Wu, Xinyi Liu, Yansheng Li
Summary: In this study, we proposed a novel method to detect marine oil spills by constructing a multi-source knowledge graph. Our method effectively organizes and utilizes various oil spill-related information and selects favorable features for oil spill detection. By combining rule inference and graph neural network technology, our method shows high sensitivity, specificity, and precision in identifying oil spills even in severely imbalanced data conditions.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2023)
Article
Remote Sensing
Wangshan Yang, Xinyi Liu, Yongjun Zhang, Yi Wan, Zheng Ji
Summary: Building instance segmentation is crucial for parallel reconstruction, management, and analysis of building instances. Existing studies have mainly focused on building scenes with large building spacing, leading to low accuracy in complex building scenes and building point clouds. To address this, we propose a novel object-based building instance segmentation method using airborne LiDAR point clouds. The method divides point clouds into objects, classifies them based on roof plane, roof accessory, and building facade characteristics, and merges objects to obtain building instances.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2022)
Article
Computer Science, Artificial Intelligence
Yongxiang Yao, Yongjun Zhang, Yi Wan, Xinyi Liu, Xiaohu Yan, Jiayuan Li
Summary: Traditional image feature matching methods are not satisfactory for multi-modal remote sensing images due to nonlinear radiation distortion differences and complicated geometric distortion. This paper proposes a new robust MRSI matching method based on co-occurrence filter space matching, which optimizes the matching by constructing a new co-occurrence scale space, extracting feature points, and optimizing the distance function. Experimental results show that the proposed method significantly outperforms other state-of-the-art methods in terms of matching effectiveness.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2022)