Article
Remote Sensing
Martin Mokros, Tomas Mikita, Arunima Singh, Julian Tomastik, Juliana Chuda, Piotr Wezyk, Karel Kuzelka, Peter Surovy, Martin Klimanek, Karolina Zieba-Kulawik, Rogerio Bobrowski, Xinlian Liang
Summary: The development of devices capable of generating 3D point clouds of the forest has flourished in recent years. Low-cost technologies such as MultiCam, iPad Pro, GeoSlam Horizon, and FARO Focus s70 were compared for tree detection and diameter at breast height estimation. Results showed that TLS provided the most accurate data, while iPad Pro achieved results closest to TLS when DBH > 7 cm.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2021)
Article
Remote Sensing
Daniel Kuekenbrink, Mauro Marty, Ruedi Boesch, Christian Ginzler
Summary: This study evaluates the performance of different close-range remote sensing devices for tree detection and diameter at breast height (DBH) extraction in forests. The results show that terrestrial laser scanning systems (TLS) have the highest tree detection rate, while drone-based laser scanning systems (UAVLS) have the lowest tree detection rate. The novel GoPro approach performs moderately well in tree detection and is comparable to LiDAR devices in DBH extraction.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2022)
Article
Remote Sensing
Xin Xu, Federico Iuricich, Kim Calders, John Armston, Leila De Floriani
Summary: This paper introduces an automated tree segmentation method based on terrestrial laser scanning technology, which utilizes a new topological algorithm to segment point clouds into individual trees without user interactions. The algorithm identifies tree bottoms and tops and reconstructs single trees using relevant topological features. Experimental results demonstrate its higher segmentation accuracy and potential for wide applications in the forest ecology community.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2023)
Article
Environmental Sciences
Sean Krisanski, Mohammad Sadegh Taskhiri, Susana Gonzalez Aracil, David Herries, Paul Turner
Summary: Forest inventories are crucial for informed decision-making in forest resource management and conservation, but the laborious process of data collection using traditional tools can lead to errors in complex forests. This research demonstrates an effective deep learning approach for semantically segmenting high-resolution forest point clouds in diverse environments, addressing the challenge of inaccurate measurements in complex forests.
Article
Forestry
ChiUng Ko, JooWon Lee, Donggeun Kim, JinTaek Kang
Summary: This study assessed the feasibility of using LiDAR devices for obtaining digital forest resource information. The findings showed that the BPLS and TLS methods had high accuracy for estimating height and DBH in most sample plots, but the BPLS underestimated height more in a sloped plot. However, the BPLS had a higher efficiency compared to the TLS method.
Article
Chemistry, Multidisciplinary
Mingrui Dai, Guohua Li
Summary: As the usage of three-dimensional (3D) laser scanner becomes prevalent for forest inventory, the analysis and processing of point cloud data captured with a 3D laser scanner have gained significant research attention. Extracting individual trees from point cloud data is crucial for further investigation at the tree-level analysis, such as tree counting and trunk analysis, leading to multiple advancements in this area. However, accurately and automatically obtaining the tree crown silhouette from the point cloud data is challenging due to the frequent overlap between adjacent tree crowns. To address this issue, a soft segmentation method utilizing K-Nearest Neighbor (KNN) and contour shape constraints has been proposed, resulting in improved visual effect and precision of point cloud segmentation. In conclusion, the proposed method demonstrates a successful approach for tree crown segmentation and silhouette reconstruction from terrestrial laser scanning point cloud data of forests.
APPLIED SCIENCES-BASEL
(2023)
Article
Plant Sciences
Peter B. Boucher, Ian Paynter, David A. Orwig, Ilan Valencius, Crystal Schaaf
Summary: The research evaluated the impact of occlusion on TLS scans and compared different stem sets, finding that occlusion from non-stem sources was the major influence on TLS line of sight. It was also discovered that transect and point TLS samples demonstrated better representativeness of some stem properties. Deriving sampled area from TLS scans improved estimates of stem density.
Article
Geography, Physical
Xufei Wang, Zexin Yang, Xiaojun Cheng, Jantien Stoter, Wenbing Xu, Zhenlun Wu, Liangliang Nan
Summary: In this research, an automatic, robust, and efficient method for registering forest point clouds is proposed. The approach locates tree stems and matches them based on their relative spatial relationship to determine the registration transformation. The algorithm requires no extra tree attributes and can align point clouds of large forest environments. Additionally, a new benchmark dataset is introduced for the development and evaluation of forest point cloud registration methods.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2023)
Article
Environmental Sciences
Zhouxin Xi, Laura Chasmer, Chris Hopkinson
Summary: This study investigates the fuel distribution in mountain pine beetle-impacted forests using terrestrial laser scanning technology. The point clouds obtained from TLS are classified into different categories using a deep learning classifier, and a quantitative structural model is used to extract volume attributes of trees and branches. The results are partially validated through field measurements, and it is found that ladder trees increase the vertical overlap between tree branches and below-canopy branches.
Article
Environmental Sciences
Carli J. Morgan, Matthew Powers, Bogdan M. Strimbu
Summary: Traditional inventories can be resource-intensive and require a trained workforce, but the use of handheld LiDAR and SfM algorithms show potential for efficient tree detection and measurement of dimensions and characteristics, such as defects and damages.
Review
Remote Sensing
Wen Xiao, Hui Cao, Miao Tang, Zhenchao Zhang, Nengcheng Chen
Summary: In recent decades, there has been an increasing focus on change detection in remote and close-range sensing due to its importance in environment monitoring and database updating. With advancements in sensing technologies, data acquisition has become more accessible and affordable, resulting in a wealth of data from various sensing platforms. This paper provides a comprehensive review of the latest developments in urban object change detection using point cloud data. It analyzes four types of objects - buildings, street scenes, urban trees, and construction sites - and summarizes the use of different data sources and open-source data with change labels. The paper thoroughly reviews change detection methods at various levels and discusses the challenges and opportunities brought by point cloud data and new techniques for future considerations.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2023)
Article
Environmental Sciences
Carine Klauberg, Jason Vogel, Ricardo Dalagnol, Matheus Pinheiro Ferreira, Caio Hamamura, Eben Broadbent, Carlos Alberto Silva
Summary: This study evaluated the damage severity of hurricanes on pine forests using terrestrial laser scanning and deep learning techniques. The combination of multicolored-by-height TLS-derived 2D images with the VGG16 convolutional neural network yielded the highest accuracy for classifying post-hurricane damage severity at the individual tree level. The researchers developed a new open-source R package (rTLsDeep) to implement all the tested methods.
Article
Environmental Sciences
Zhouxin Xi, Chris Hopkinson
Summary: The treeiso model enables accurate isolation of individual trees from TLS scans, providing strong support for fine-scale forest management. Sensitivity analysis and parameter combination testing demonstrate the model's good accuracy and robustness in tree localization.
Article
Geography, Physical
Luigi Parente, Jim H. Chandler, Neil Dixon
Summary: This paper evaluates an automated registration strategy using the SIFT algorithm, known as Time-SIFT, and found that it can achieve accurate multitemporal point cloud alignments under various conditions. Combining the Time-SIFT approach with the ICP algorithm results in moderate alignment improvements, and the use of innovative direct georeferencing technique allows for accurate georectification of 3D models.
PHOTOGRAMMETRIC RECORD
(2021)
Article
Geography, Physical
Luigi Parente, Jim H. Chandler, Neil Dixon
Summary: The paper assesses an automated registration strategy using the scale-invariant feature transform (SIFT) algorithm in modern photogrammetric software, demonstrating accurate alignment of multitemporal point clouds. The combination of the Time-SIFT approach with an ICP algorithm showed moderate improvements in alignment accuracy. Additionally, the use of an innovative direct georeferencing technique with a robotic total station enabled accurate georectification of 3D models.
PHOTOGRAMMETRIC RECORD
(2021)
Article
Computer Science, Theory & Methods
Sudheer Kumar Battula, Malgorzata M. OaReilly, Saurabh Garg, James Montgomery
Summary: This study provides a continuous-time Markov chain (CTMC) based resource availability model for Fog computing environments, demonstrating its applicability by integrating it with the Best-Fit resource selection policies.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2021)
Article
Forestry
Martin Strandgard, Mohammad Sadegh Taskhiri, Mauricio Acuna, Paul Turner
Summary: Australia's potential forest bioenergy resource is largely underutilized due to high delivery costs. Drying forest biomass at the roadside and storing it with high spatial density can significantly reduce delivered costs by about 60%. Whole trees were found to be the most cost-effective resource, while conventional storage methods for LR may increase costs.
Article
Environmental Sciences
Sean Krisanski, Mohammad Sadegh Taskhiri, Susana Gonzalez Aracil, David Herries, Paul Turner
Summary: Forest inventories are crucial for informed decision-making in forest resource management and conservation, but the laborious process of data collection using traditional tools can lead to errors in complex forests. This research demonstrates an effective deep learning approach for semantically segmenting high-resolution forest point clouds in diverse environments, addressing the challenge of inaccurate measurements in complex forests.
Article
Computer Science, Interdisciplinary Applications
Mihai Neagoe, Hans-Henrik Hvolby, Mohammad Sadegh Taskhiri, Paul Turner
Summary: This research compares the impact of infrastructure congestion management initiatives at a bulk cargo marine terminal on truck queuing and emissions. Results show that terminal appointment systems are one of the most effective congestion mitigation initiatives, reducing truck turnaround times by up to 65% and engine idling emissions by up to 80%. The study highlights the differential impacts of congestion management initiatives on truck and environmental performance in bulk cargo marine terminals.
SIMULATION MODELLING PRACTICE AND THEORY
(2021)
Article
Forestry
Martin Strandgard, Paul Turner, Anna Shillabeer
Summary: Forest biomass could be a significant energy source in Australia, but reducing costs is crucial. This paper introduces an operational decision support system for scheduling biomass deliveries at the lowest cost, taking into account moisture changes. By comparing different mathematical models, a Greedy algorithm is found to be the best performing in terms of speed and cost, and can be easily modified to improve implementation. The accuracy of stored biomass quantity estimates is identified as a critical area for future research.
Review
Forestry
Jamal Maktoubian, Mohammad Sadegh Taskhiri, Paul Turner
Summary: This paper discusses the impact of machinery maintenance and machine learning on improving the reliability of bioenergy supply chains from woody biomass, highlighting some challenges that still exist in the field of forestry machinery maintenance. It emphasizes the potential of big data analytics in enhancing the identification and prediction of maintenance needs, as well as the lack of research on the effects of external factors on maintenance costs.
Article
Forestry
Sean Krisanski, Mohammad Sadegh Taskhiri, James Montgomery, Paul Turner
Summary: Unoccupied Aircraft Systems (UAS) are being used as a replacement for conventional forest plot mensuration to monitor growth, estimate biomass, evaluate carbon stocks, and detect weeds. The use of UAS can improve safety, efficiency, and reduce costs compared to traditional sampling techniques. This study describes and demonstrates four iterations of 3D printed canopy sampling UAS, with the fourth iteration successfully collecting canopy samples from three tree species. Further development is envisioned to integrate advanced remote sensing techniques and achieve a fully-automated forest information capture system.
Review
Chemistry, Multidisciplinary
Zhaoyuan He, Paul Turner
Summary: This paper provides a systematic review of blockchain applications in the forestry sector, highlighting their benefits, opportunities, and challenges. The study reveals that blockchain has great potential in forestry, but it is still immature and complex.
APPLIED SCIENCES-BASEL
(2022)
Article
Chemistry, Multidisciplinary
Saleem Ameen, Ming-Chao Wong, Kwang-Chien Yee, Paul Turner
Summary: Advances in AI in healthcare have the potential to improve clinical decisions and patient care, but the reductive reasoning and computational determinism used in AI systems pose limitations and risks. The paper highlights the biases in AI systems, the limited evaluations conducted, and the marginalization of socio-technical factors in CRC.
APPLIED SCIENCES-BASEL
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Stephen Chen, Antonio Bolufe-Rohler, James Montgomery, Dania Tamayo-Vera, Tim Hendtlass
Summary: The rate of successful exploration is influenced by the proportion of fitter solutions in the attraction basins and how the reference solution moves towards its local optimum. Increased exploitation of the reference solution can lead to a decrease in the proportion of fitter solutions and an increase in failed exploration rates. This effect is studied in Particle Swarm Optimization and Differential Evolution algorithms, highlighting the importance of addressing the curse of dimensionality in metaheuristic design.
2022 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC)
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Stephen Chen, Antonio Bolufe-Rohler, James Montgomery, Wenxuan Zhang, Tim Hendtlass
Summary: Metaheuristics for numerical optimization tend to decline in performance as dimensionality increases, known as the Curse of Dimensionality. The volume of search spaces grows exponentially with increasing dimensionality, leading to extensive research on improved exploration. Recent insights show that attraction basins can also change significantly, leading to selection-based approaches to combat the Curse of Dimensionality. Introduction of Average-Fitness Based Selection reduces the rate of selection errors with increasing dimensionality.
2022 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC)
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Stephen Chen, Shehnaz Islam, Antonio Bolufe-Rohler, James Montgomery, Tim Hendtlass
Summary: Random walks are useful for modeling stochastic processes and can provide insights into practical situations with added model features. By conducting experiments on random walks, the effects of selection, exploration, and exploitation during metaheuristics search processes were studied. The experiments showed that movement increases with higher dimensionality of sampling distributions, while Simulated Annealing has improved performance with increasing dimensionality compared to Particle Swarm Optimization.
2021 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC 2021)
(2021)