Article
Environmental Sciences
Xiaoyan Zhang, Linhui Li, Donglin Di, Jian Wang, Guangsheng Chen, Weipeng Jing, Mahmoud Emam
Summary: In this paper, we propose an improved squeeze and excitation residual network (SERNet) for the semantic segmentation of high-resolution remote sensing images. The SERNet integrates several squeeze and excitation residual modules (SERMs) and a refine attention module (RAM), which effectively addresses the challenges posed by complex distribution of ground objects and unclear boundaries. Experimental results demonstrate the superior performance of SERNet on the ISPRS datasets.
Article
Plant Sciences
Jian Lin, Qin Ma, Yang Ju, Hongsheng Zhang, Qiang Wang, Bo Huang
Summary: Urban trees store and sequester large amounts of carbon and play a crucial role in natural climate solutions. However, there is limited understanding of how the spatial distribution of carbon density varies across different social, demographic, and built dimensions of urban landscapes. Moreover, the impact of landscape structure and design on carbon densities in urban trees remains unclear.
URBAN FORESTRY & URBAN GREENING
(2022)
Article
Geochemistry & Geophysics
Yizhang Liu, Yanping Li, Luanyuan Dai, Taotao Lai, Changcai Yang, Lifang Wei, Riqing Chen
Summary: This method achieves the best performance by integrating motion consistency into the general region growing pipeline to handle different deformations.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Geochemistry & Geophysics
Yonghao Xu, Pedram Ghamisi
Summary: This study systematically analyzes the universal adversarial examples in remote sensing data, and proposes a new black-box adversarial attack method that can generate transferable adversarial examples.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Ally Kruper, Robert J. McGaughey, Sarah Crumrine, Bernard T. Bormann, Keven Bennett, Courtney R. Bobsin
Summary: This study combined LiDAR data with field data to improve the accuracy of tree coordinates and differentiate between specific tree species.
Article
Forestry
Guido Ceccherini, Gregory Duveiller, Giacomo Grassi, Guido Lemoine, Valerio Avitabile, Roberto Pilli, Alessandro Cescatti
Summary: This article discusses the importance of timely and accurate monitoring of forest resources, given the multifunctionality of these ecosystems and their increasing vulnerability to climate change. It analyzes the strengths and weaknesses of remote sensing observations and ground observations from National Forest Inventories (NFIs) as the two major sources of information for assessing forest area and use. The integration of satellite and surface data is proposed as a novel method to address the limitations of the current approaches. The article also highlights the need to collect in situ data that is relevant and compatible with remote sensing products within the European Union.
ANNALS OF FOREST SCIENCE
(2022)
Article
Environmental Sciences
Heather North, Alexander Amies, John Dymond, Stella Belliss, David Pairman, John Drewry, Jan Schindler, James Shepherd
Summary: This study introduces an automated mapping method using time-series satellite imagery to map bare ground extent in hill-country forestry and winter forage grazing land in New Zealand, providing national maps and statistics for 2018 and assessing accuracy.
JOURNAL OF ENVIRONMENTAL MANAGEMENT
(2022)
Article
Environmental Sciences
Ana Ri, Huijun An
Summary: This study proposes a more accurate assessment method for forest health in the natural larch pine forests of Arxan by integrating remote sensing technology with tree crown feature analysis. It introduces deep learning technology and a spectral-Gabor space discrimination and classification model to analyze multi-spectral remote sensing image features. The incorporation of quantitative indicators, such as tree crown features, enhances the accuracy and efficiency of forest health assessment.
FRONTIERS IN ENVIRONMENTAL SCIENCE
(2023)
Article
Geochemistry & Geophysics
Gencer Sumbul, Mahdyar Ravanbakhsh, Beguem Demir
Summary: This article proposes a novel triplet sampling method in the framework of deep neural networks for multilabel remote sensing image retrieval. The method selects a small set of representative and informative triplets to reduce computational complexity and improve learning speed.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Review
Forestry
Riccardo Dainelli, Piero Toscano, Salvatore Filippo Di Gennaro, Alessandro Matese
Summary: The study systematically analyzes the application of UAVs in forestry ecosystems, aiming to create a reference dataset for researchers and identify trends and gaps in research. Challenges include promoting UAV-RS in tropical and equatorial forests and further research on hyperspectral sensors for long-term monitoring.
Article
Geochemistry & Geophysics
Haifeng Wang, Hualong Cao, Yan Kai, Haicheng Bai, Xuefeng Chen, Yang Yang, Lin Xing, Chengjiang Zhou
Summary: This study proposes an object detection network based on the dynamic extraction of multisource image features to quickly and accurately locate damaged objects in disaster areas. By collecting multisource remote-sensing images before and after the disaster for training, the accuracy of the detection model in this study exceeds 85% when the detection error rate is less than 5%.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Geochemistry & Geophysics
Xiangli Nie, Ruofei Gao, Rui Wang, Deliang Xiang
Summary: A novel online multiview deep forest architecture is proposed in this study, which processes data from different views through a cascade structure and multiple layers of ensemble, resulting in learning a deep forest model in an online manner from a stream of multiview data.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2021)
Review
Environmental Sciences
Svetlana Illarionova, Dmitrii Shadrin, Polina Tregubova, Vladimir Ignatiev, Albert Efimov, Ivan Oseledets, Evgeny Burnaev
Summary: This article discusses the application of computer vision techniques in forest management actions. By utilizing remote sensing data, tasks such as estimation of forest areas, tree species classification, and estimation of forest resources can be achieved. The implementation of these techniques is crucial for improving forest management strategies, implementing carbon offset projects, and enhancing the accuracy of system change predictions.
Review
Forestry
Jaz Stoddart, Juan Suarez, William Mason, Ruben Valbuena
Summary: Continuous cover forestry is a sustainable management approach that requires new evaluation methods and solutions to challenges through the application of remote sensing techniques.
CURRENT FORESTRY REPORTS
(2023)
Article
Engineering, Electrical & Electronic
Songyu Zhu, Weipeng Jing, Peilun Kang, Mahmoud Emam, Chao Li
Summary: In this study, a forest remote sensing change detection model in the context of few-shot learning is proposed. The proposed model achieves end-to-end change detection algorithm for forest scenes through data augmentation and updated few-shot algorithm. The model improves the F1 score from 86% to 91% on the two datasets, and increases the F1 score by 6.52% on average.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2023)