Article
Automation & Control Systems
Zuowei Zhang, Songtao Ye, Yiru Zhang, Weiping Ding, Hao Wang
Summary: This paper proposes a new classifier method based on evidence theory to handle missing values, improving classification performance by addressing the uncertainty and imprecision brought by incompleteness. Experimental results demonstrate that the proposed method outperforms other methods in terms of accuracy, precision, recall, and F1 measure, but with higher computational costs.
IEEE-CAA JOURNAL OF AUTOMATICA SINICA
(2022)
Article
Automation & Control Systems
Jiawei Niu, Zhunga Liu, Yao Lu, Zaidao Wen
Summary: The paper introduces a new method called evidential combination of classifiers (ECC) for handling imbalanced data by combining hybrid-sampling, over-sampling, and under-sampling methods. By revising the classification results at the decision level, the method aims to reduce error risk and achieve improved classification performance through evidence theory combination.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2022)
Article
Chemistry, Analytical
Sang-hyub Lee, Deok-Won Lee, Kooksung Jun, Wonjun Lee, Mun Sang Kim
Summary: A new method for combining multiple inaccurate skeleton data sets from different sensors into a single accurate skeleton data was proposed in this paper, using DBSCAN and Kalman filter for noise and error reduction. Increasing the number of sensors improved the joint position accuracy of the merged skeleton, with the best performance shown when the DBSCAN searching area was 10 cm.
Article
Agriculture, Multidisciplinary
John Bonestroo, Mariska van der Voort, Henk Hogeveen, Ulf Emanuelson, Ilka Christine Klaas, Nils Fall
Summary: This study developed a sensor-based prediction model using machine learning to forecast the chronicity in subclinical mastitis cases, showing high accuracy in predicting future chronic mastitis status based on past sensor data.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2022)
Article
Automation & Control Systems
Virginie Felizardo, Nuno M. Garcia, Imen Megdiche, Nuno Pombo, Miguel Sousa, Frantisek Babic
Summary: The need to balance insulin, food, and exercise in controlling diabetes creates an opportunity for developing mobile applications for self-management. This study proposes a hypoglycaemia prediction approach using information fusion and classifiers consensus, which shows promising results in predicting the risk of hypoglycaemia within a 24-hour window.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Automation & Control Systems
Diego Galvan, Jelmir Craveiro de Andrade, Carlos Adam Conte-Junior, Mario Henrique M. Killner, Evandro Bona
Summary: This study proposes a new method for authentication that combines fingerprint analytical techniques with multivariate methods of one-class classifiers. The method utilizes common dimension analysis and dual data-driven to calculate misclassification errors and determine their cut-off levels. The results demonstrate that this method improves the quality and efficiency of the authentication model compared to traditional methods, with the mid-level data fusion approach providing clearer separation between target and non-target classes. Further applications are being conducted to confirm the reliability of the method with other matrices and analytical techniques.
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS
(2023)
Article
Engineering, Electrical & Electronic
Bruno Rodrigues, Eder J. Scheid, Julius Willems, Maximilian Tornow, Katharina O. E. Muller, Burkhard Stiller
Summary: This article presents a method called FusIon Data Tracking System (FITS) as an approach and proof-of-concept to correlate data from different indoor sensors to movement profiles of different individuals. FITS achieves this by generating synthetic sensor measurement data and effectively solving clustering and position prediction tasks, improving the accuracy and effectiveness of indoor tracking.
IEEE SENSORS JOURNAL
(2022)
Article
Engineering, Electrical & Electronic
Hongkang Wei, Li Ma, Yong Liu, Qian Du
Summary: This article investigates the effectiveness of multiclassifier fusion technique on domain adaptation for remote sensing image classification. It proposes the multiple domain adaptation fusion (MDAF) method and the multiple base classifier fusion (MBCF) method, showing the efficiency of the proposed weighting strategy in these methods through experiments with three remote sensing images.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2021)
Article
Robotics
Alex Wong, Safa Cicek, Stefano Soatto
Summary: This method leverages synthetic data to learn the association of sparse point clouds with dense natural shapes and uses images as evidence to validate depth maps, achieving state-of-the-art performance on both indoor and outdoor benchmark datasets.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2021)
Review
Computer Science, Artificial Intelligence
Chahinez Ounoughi, Sadok Ben Yahia
Summary: This paper presents a systematic literature review on recent data fusion methods in intelligent transportation systems (ITS) and identifies the main challenges and issues. It describes the characteristics of current data fusion methods using multi-sensor heterogeneous data sources and highlights research gaps, current challenges, and new research trends.
INFORMATION FUSION
(2023)
Article
Computer Science, Information Systems
Julian Webber, Abolfazl Mehbodniya, Ahmed Arafa, Ahmed Alwakeel
Summary: Human activity recognition is widely used in healthcare, daily life, and security fields. Combining classifiers with different complexities and multiple sensors can improve classification accuracy.
Article
Robotics
Ran Long, Christian Rauch, Tianwei Zhang, Vladimir Ivan, Tin Lun Lam, Sethu Vijayakumar
Summary: This work presents a novel dense RGB-D SLAM approach for dynamic planar environments that enables simultaneous multi-object tracking, camera localisation, and background reconstruction. The approach utilizes camera motion prior to perform dense SLAM even when the camera view is largely occluded by multiple dynamic objects.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2022)
Article
Construction & Building Technology
Jingyuan Tang, Mingzhu Wang, Han Luo, Peter Kok-Yiu Wong, Xiao Zhang, Weiwei Chen, Jack C. P. Cheng
Summary: To reduce accidents involving heavy machines on construction sites, it is crucial to automatically monitor the full-body poses of the operating machines. Conventional pose estimation systems using homogeneous sensors are vulnerable to negative environmental impacts, resulting in inaccurate and unstable estimation of machine states. Hence, a full-body pose estimation framework is proposed for excavators, utilizing a data fusion strategy that incorporates different types of onboard sensors to enhance accuracy and robustness. Through competitive and complementary data fusion, key-points describing the full-body poses of the excavator are tracked in 3D space, improving the accuracy and robustness of pose estimation. Especially, an EKF-based localization algorithm is developed for optimized multi-keypoint tracking and validated through a real-world excavator case study. The proposed sensor-fusion method effectively enhances operational safety by accurately monitoring the motion of heavy machines on construction sites.
AUTOMATION IN CONSTRUCTION
(2023)
Article
Chemistry, Analytical
Zhaofeng Niu, Yuichiro Fujimoto, Masayuki Kanbara, Taishi Sawabe, Hirokazu Kato
Summary: In this paper, we propose a new TSDF fusion network called DFusion, which aims to minimize the influences of depth noises and pose noises on the 3D reconstruction process. The network consists of a fusion module that generates a TSDF volume and a denoising module that removes both depth and pose noises. 3D convolutional layers and a specially-designed loss function are utilized to utilize the 3D structural information of the TSDF volume and improve the fusion performance. Experimental results demonstrate the superiority of our method over existing methods.
Article
Engineering, Electrical & Electronic
Elnaz Namazi, Rudolf Mester, Jingyue Li, Chaoru Lu, Meng Tang, Ying Xiong
Summary: This paper investigates a new methodology for estimating and fusing real-time traffic data using multiple vehicles equipped with low-cost sensors in an intelligent traffic management system. The research shows that the proposed method significantly improves the accuracy of geolocation estimation for target vehicles.
IEEE SENSORS JOURNAL
(2022)
Article
Remote Sensing
Longwei Li, Nan Li, Zhuo Zang, Dengsheng Lu, Guangxing Wang, Ni Wang
Summary: Moso bamboo has unique characteristics such as fast growth rate, short harvesting cycle, and on/off-year phenomenon. This research used data from the VEN mu S micro-satellite to analyze the phenological features of Moso bamboo forests, determining sensitive spectral ranges for seasonal variation and identifying different phenological periods using the Red-edge Position Index (REPI).
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2021)
Article
Geography, Physical
Yaoliang Chen, Xiaotao Huang, Jingfeng Huang, Shanshan Liu, Dengsheng Lu, Shuai Zhao
Summary: The study successfully monitored the dynamics of desert vegetation in a dryland basin of Northwest China using MESMA method. Results showed areas of degradation, recovery, and greening, and analyzed different influencing factors, demonstrating the potential application of this method in semi-arid and arid regions.
GISCIENCE & REMOTE SENSING
(2021)
Article
Biodiversity Conservation
Dengqiu Li, Dengsheng Lu, Yan Zhao, Mingxing Zhou, Guangsheng Chen
Summary: This study developed a new framework to assess habitat fragmentation by integrating spatial patterns of vegetation coverage and its change in the giant panda habitat ecosystem of China. The results revealed that most disturbed areas experienced negative abrupt vegetation change, while undisturbed areas mainly showed an increase in vegetation. By identifying spatial clusters and outliers, the study pinpointed areas in need of careful management to reduce habitat fragmentation.
ECOLOGICAL INDICATORS
(2021)
Article
Computer Science, Interdisciplinary Applications
James D. A. Millington, Valeri Katerinchuk, Ramon Felipe Bicudo da Silva, Daniel de Castro Victoria, Mateus Batistella
Summary: The increasing global demand for agricultural commodities has driven local land use/cover change and agricultural production in Brazil during the 21st century. CRAFTY-Brazil is a LUCC model representing the production of multiple agricultural commodities, taking into account spatially explicit and temporally contingent processes in the nearly four million km2 Brazilian study area.
ENVIRONMENTAL MODELLING & SOFTWARE
(2021)
Article
Multidisciplinary Sciences
Ramon Felipe Bicudo da Silva, Andres Vina, Emilio F. Moran, Yue Dou, Mateus Batistella, Jianguo Liu
Summary: Human-environment interactions across borders are now more influential than ever, posing unprecedented sustainability challenges. The metacoupling framework provides a useful tool to evaluate these interactions at diverse temporal and spatial scales.
SCIENTIFIC REPORTS
(2021)
Article
Multidisciplinary Sciences
Daniel de Castro Victoria, Ramon Felipe Bicudo da Silva, James D. A. Millington, Valeri Katerinchuk, Mateus Batistella
Summary: Transport costs play a significant role in agricultural exports, business profitability, and farmers' decisions on land use. This study evaluated the impacts of changing global demand for agricultural products on land use and cover change using the CRAFTY-Brazil model. The research reconstructed the historic road network and identified export ports to derive transport costs for Brazil in different years.
Article
Geography, Physical
Wenke Lin, Yagang Lu, Guiying Li, Xiandie Jiang, Dengsheng Lu
Summary: This study compares the performance of LS-CHM and L-CHM for FGSV modeling and explores the advantages of using the hierarchical Bayesian approach when sample size is small. The results show that L-CHM provides better predictions overall using the same modeling approaches, but LS-CHM-based variables produce better modeling accuracy than L-CHM-based variables in a specific range of FGSV. The HBA based on stratification of both forest type and slope aspect provides the best FGSV estimation.
GISCIENCE & REMOTE SENSING
(2022)
Article
Environmental Sciences
Mengzhuo Fan, Kuo Liao, Dengsheng Lu, Dengqiu Li
Summary: Examining the characteristics and spatial patterns of vegetation change under different protection levels can provide a scientific basis for national park protection and management. The study analyzed the vegetation change in Wuyishan National Park using Landsat EVI data from 1986 to 2020 and the WBS approach. The results showed that the highest percentage of area without abrupt change was in the strictly protected area, while the non-protected area had the lowest percentage. The study also found that the vegetation coverage generally improved in the park, with higher positive percentage in the protected areas. However, the non-protected area had a higher mean greenness change. The study highlighted the importance of protection level in determining vegetation change and spatial patterns in the national park.
Article
Environmental Sciences
Kuo Liao, Yunhe Li, Bingzhang Zou, Dengqiu Li, Dengsheng Lu
Summary: This study compared the accuracy of tree height measurements using different methods and the influence of allometric models on tree volume estimation accuracy. The results showed significant impacts of different measurement methods on tree volume calculations, and incorporating UAV Lidar data with DBH field measurements can effectively improve tree volume estimation accuracy.
Article
Environmental Sciences
Yi Zhang, Dengsheng Lu, Xiandie Jiang, Yunhe Li, Dengqiu Li
Summary: In this study, the 3-PG model was optimized and calibrated using survey and UAV lidar data at the sample plot scale and applied at the forest sub-compartment scale. The results show that both survey forests age data and remote-sensing-derived forest age data can accurately estimate eucalyptus plantation parameters. The simulation results based on remote-sensed forest age data are significantly better than the ones based on survey data, providing an important reference for future studies using remote sensing-derived forest age data in large spatial scales.
Article
Remote Sensing
Ruoqi Wang, Guiying Li, Yagang Lu, Dengsheng Lu
Summary: This research compared the advantages of using object-based GSV modeling approach with traditional grid-based approaches for poplar GSV estimation. The results showed that the object-based approach was more accurate in estimating GSV and solving the mixed plot problem in the striped forest distribution areas.
GEO-SPATIAL INFORMATION SCIENCE
(2023)
Article
Environmental Sciences
Yongpeng Ye, Dengsheng Lu, Zuohang Wu, Kuo Liao, Mingxing Zhou, Kai Jian, Dengqiu Li
Summary: This study developed a framework based on multi-source high-resolution satellite images to analyze vertical characteristics of mountainous vegetation distribution. The results showed distinct differentiation of vegetation types along elevation gradients in Wuyishan National Park, with significant differences in distribution patterns under different human protection levels.
Article
Remote Sensing
Willian Vieira de Oliveira, Luciano Vieira Dutra, Sidnei Joao Siqueira Sant'Anna
Summary: In land-cover classification using remote sensing images, methods that solely analyze the spectral information of individual pixels often generate noisy results. Incorporating spatial contextual information into classification can effectively reduce noise and improve accuracy. However, existing contextual methods may oversmooth certain classes, causing the loss of important spatial structures. To address this issue, a strategy called Meta-CTX is proposed, which allows for a trade-off between noise smoothing and the preservation of small spatial details.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2022)
Article
Ecology
Lucia Zarba, Maria Piquer-Rodriguez, Sebastien Boillat, Christian Levers, Ignacio Gasparri, T. Mitchell Aide, Nora L. Alvarez-Berrios, Liana O. Anderson, Ezequiel Araoz, Eugenio Arima, Mateus Batistella, Marco Calderon-Loor, Cristian Echeverria, Mariano Gonzalez-Roglich, Esteban G. Jobbagy, Sarah -Lan Mathez-Stiefel, Carlos Ramirez-Reyes, Andrea Pacheco, Maria Vallejos, Kenneth R. Young, Ricardo Grau
Summary: This study presents a spatially explicit social-ecological land system (SELS) typology for South America, which incorporates both social and biophysical dimensions. Through a hybrid methodology combining data-driven spatial analysis and knowledge-based evaluation by regional specialists, the study identifies 13 SELS in 5 larger social-ecological regions. The SELS classification provides a systematic and operative characterization of South American social-ecological land systems.
ECOLOGY AND SOCIETY
(2022)
Article
Forestry
Shiyun Pang, Guiying Li, Xiandie Jiang, Yaoliang Chen, Yagang Lu, Dengsheng Lu
Summary: This research aims to explore the method of accurately retrieving forest canopy height from ATLAS data and improve retrieval accuracy by incorporating a high-precision digital terrain model (DTM) and a data-filtering strategy. The results show that using the proposed method, the retrieval accuracy of forest canopy height in mountainous regions with dense forest cover and complex terrain conditions can be considerably improved.