Zhizhong Kang

ORCID: 0000-0002-9728-4702
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Remote Sensing and LiDAR Applications
  • 3D Surveying and Cultural Heritage
  • Robotics and Sensor-Based Localization
  • Planetary Science and Exploration
  • Astro and Planetary Science
  • Advanced Image and Video Retrieval Techniques
  • Thermochemical Biomass Conversion Processes
  • 3D Modeling in Geospatial Applications
  • Advanced Vision and Imaging
  • Coal Properties and Utilization
  • Coal and Its By-products
  • Image Processing and 3D Reconstruction
  • Image and Object Detection Techniques
  • Space Science and Extraterrestrial Life
  • Cryospheric studies and observations
  • Remote Sensing in Agriculture
  • Geology and Paleoclimatology Research
  • Advanced Neural Network Applications
  • Geochemistry and Geologic Mapping
  • Combustion and flame dynamics
  • Computer Graphics and Visualization Techniques
  • Automated Road and Building Extraction
  • Space Exploration and Technology
  • Inertial Sensor and Navigation
  • Geographic Information Systems Studies

China University of Geosciences (Beijing)
2016-2025

Harbin Institute of Technology
2025

Jiangnan University
2024

Ministry of Education of the People's Republic of China
2020-2024

Sun Yat-sen University
2024

North China Electric Power University
2010-2024

Third Affiliated Hospital of Sun Yat-sen University
2024

Ministry of Natural Resources
2024

Capital Normal University
2018

Hunan University
2015

Accurate individual tree segmentation is an important basis for the subsequent calculation and analysis of forestry parameters. However, rasterized canopy height model based methods often suffer from 3-D information loss due to interpolation operation. Therefore, this article proposes method on marker-controlled watershed algorithm spatial distribution airborne LiDAR point clouds. First, potential apices derived local maxima filtering, conducted obtain coarse clusters. Then, within principal...

10.1109/jstars.2020.2979369 article EN cc-by IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 2020-01-01

Abstract. Automated generation of 3D indoor models from point cloud data has been a topic intensive research in recent years. While results on various datasets have reported literature, comparison the performance different methods not possible due to lack benchmark and common evaluation framework. The ISPRS modelling aims address this issue by providing public dataset an framework for methods. In paper, we present comprising several clouds environments captured sensors. We also discuss based...

10.5194/isprs-archives-xlii-2-w7-367-2017 article EN cc-by ˜The œinternational archives of the photogrammetry, remote sensing and spatial information sciences/International archives of the photogrammetry, remote sensing and spatial information sciences 2017-09-12

Water leakages can affect the safety and durability of shield tunnels, so rapid accurate identification diagnosis are urgently needed. However, current leakage detection methods mostly based on mobile LiDAR data, making it challenging to detect damage in both terrestrial data simultaneously, results not intuitive. Therefore, an integrated cylindrical voxel Mask R-CNN method for water inspection is presented this paper. This includes following three steps: (1) a 3D cylindrical-voxel...

10.3390/rs16050896 article EN cc-by Remote Sensing 2024-03-03

This paper presents a new approach to the automatic registration of terrestrial laser scanning (TLS) point clouds using panoramic reflectance images. The follows two-step procedure that includes both pair-wise and global registration. consists image matching (pixel-to-pixel correspondence) cloud (point-to-point correspondence), as correspondence between (pixel-to-point) is inherent False correspondences are removed by geometric invariance check. pixel-to-point computation rigid...

10.3390/s90402621 article EN cc-by Sensors 2009-04-15

An efficient method for the continuous extraction of subway tunnel cross sections using terrestrial point clouds is proposed. First, central axis extracted a 2D projection cloud and curve fitting RANSAC (RANdom SAmple Consensus) algorithm, optimized global strategy based on segment-wise fitting. The cross-sectional planes, which are orthogonal to axis, then determined every interval. points by intersecting straight lines that rotate orthogonally around within plane with cloud. interpolation...

10.3390/rs6010857 article EN cc-by Remote Sensing 2014-01-15

The safety of the electricity infrastructure significantly affects both our daily life and industrial activities. Timely accurate monitoring network can prevent dangerous situations effectively. Thus, we, in this paper, develop a voxel-based method for automatically extracting transmission lines from airborne LiDAR point cloud data. proposed paper uses three-dimensional (3-D) voxels as primitives consist following steps: First, skeleton structure extraction using Laplacian smoothing; second,...

10.1109/jstars.2018.2869542 article EN IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 2018-09-26

In this paper, a new method for novel X-ray pulsar navigation is proposed to overcome the Doppler effects from motion of deep space explorer. An analysis was undertaken dynamic orbit model interplanetary trajectory cruise phase. During signal observation period, explorer can be considered at constant acceleration motion. A compensation based on analysis. The demonstrates great advantages in terms low computational cost. However, there an evident bias due pulse time-of-arrival (TOA)....

10.1109/taes.2014.130463 article EN IEEE Transactions on Aerospace and Electronic Systems 2015-01-01

10.1016/j.isprsjprs.2018.04.018 article EN ISPRS Journal of Photogrammetry and Remote Sensing 2018-05-24

The digital mapping of road environment is an important task for infrastructure inventory and urban planning. Automatic extraction classification pole-like objects can remarkably reduce cost enhance work efficiency. Therefore, this paper proposes a voxel-based method that automatically extracts classifies three-dimensional (3-D) by analyzing the spatial characteristics objects. First, shape recognition conducted to generate set object candidates. Second, according their isolation vertical...

10.1109/jstars.2018.2869801 article EN IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 2018-11-01

This paper presents a novel framework to extract metro tunnel cross sections (profiles) from Terrestrial Laser Scanning point clouds. The entire consists of two steps: central axis extraction and section determination. In extraction, we propose slice-based method obtain an initial axis, which is further divided into linear nonlinear circular segments by enhanced Random Sample Consensus (RANSAC) segmentation algorithm. algorithm transforms the problem hybrid segment sole elements defined at...

10.3390/rs11030297 article EN cc-by Remote Sensing 2019-02-01

Point cloud classification is of great importance to applications airborne Light Detection And Ranging (LiDAR) data. In recent years, LiDAR has been integrated with various other sensors, e.g., optical imaging and thus, the fusion multiple data types for scene become a hot topic. Therefore, this paper proposes Bayesian network (BN) model that suitable point fusing types. Based on an analysis characteristics clouds aerial images, we first extract geometric features from spectral images. The...

10.1109/jstars.2016.2628775 article EN IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 2016-12-12

This paper reports the results of ISPRS benchmark on indoor modelling. Reconstructed models submitted by 11 participating teams are evaluated a dataset comprising 6 point clouds representing environments different complexity. The evaluation is based measuring completeness, correctness, and accuracy reconstructed wall elements through comparison with manually generated reference models. show that performance methods varies across datasets, but generally reconstruction achieve better for...

10.1016/j.ophoto.2021.100008 article EN cc-by-nc-nd ISPRS Open Journal of Photogrammetry and Remote Sensing 2021-10-29

Indoor-scene semantic segmentation is of great significance to indoor navigation, high-precision map creation, route planning, etc. However, incorporating RGB and HHA images for indoor-scene a promising yet challenging task, due the diversity textures structures disparity multi-modality in physical significance. In this paper, we propose Cross-Modality Attention Network (CMANet) that facilitates extraction both features enhances cross-modality feature integration. CMANet constructed under...

10.3390/s22218520 article EN cc-by Sensors 2022-11-05

Urban scene-level 3D point cloud labeling is a very laborious and expensive task compared to images. Conversely however, image processing techniques, deep learning or otherwise are more established mature. Thus, in multi-source data environment, the of scene via an automated process as initial step, followed by manual human verification effective way save man hours cost. With above goal, this study presents simple but robust spatio-spectral feature representation approach. In approach,...

10.1016/j.jag.2023.103302 article EN cc-by-nc-nd International Journal of Applied Earth Observation and Geoinformation 2023-05-19

Mobile Laser Scanning (MLS) point cloud data contains rich three-dimensional (3D) information on road ancillary facilities such as street lamps, traffic signs and utility poles. Automatically recognizing from would provide benefits for safety inspection, management so on, can also basic support the construction of an city. This paper presents a method extracting classifying pole-like objects (PLOs) unstructured MLS data. Firstly, is preprocessed to remove outliers, downsample filter ground...

10.3390/rs10121891 article EN cc-by Remote Sensing 2018-11-27

Light Detection and Ranging (LiDAR) has advantages in detecting individual trees because it can obtain information on the vertical structure even lower layers. However, current methods still cannot detect understory well, small are often clumped together overlapped by large trees. To fill this gap, a two-stage network named Tree Region-Based Convolutional Neural Network (RCNN) was proposed to directly from point clouds. In first stage, very dense anchors generated anywhere forest. Then, RCNN...

10.3390/rs15041024 article EN cc-by Remote Sensing 2023-02-13

Abstract Impact craters are geomorphological features widely distributed on the lunar surface. Their morphological parameters crucial for studying reasons their formation, thickness of regolith at impact site and age crater. However, current research extraction multiple from a large number within extensive geographical regions faces several challenges, including issues related to coordinate offsets in heterogeneous data, insufficient interpretation crater profile morphology incomplete...

10.1111/phor.12483 article EN The Photogrammetric Record 2024-03-25

It is generally believed that employing partially coherent light for wireless optical communication can improve the performance. In this paper, we show whether partial coherence contributes positively or negatively depends on turbulence strength of link. For illustration, self-focusing vortex (PCSFV) beams propagating via anisotropic atmospheric at different altitudes are investigated. shown lower improves focusing and helps signal receiving only low-altitude strong turbulence. There an...

10.1364/josab.523505 article EN Journal of the Optical Society of America B 2024-04-29
Coming Soon ...