Article
Multidisciplinary Sciences
Radost Stanimirova, Katelyn Tarrio, Konrad Turlej, Kristina McAvoy, Sophia Stonebrook, Kai-Ting Hu, Paulo Arevalo, Eric L. Bullock, Yingtong Zhang, Curtis E. Woodcock, Pontus Olofsson, Zhe Zhu, Christopher P. Barber, Carlos M. Souza Jr, Shijuan Chen, Jonathan A. Wang, Foster Mensah, Marco Calderon-Loor, Michalis Hadjikakou, Brett A. Bryan, Jordan Graesser, Dereje L. Beyene, Brian Mutasha, Sylvester Siame, Abel Siampale, Mark A. Friedl
Summary: This study created a global database of nearly 2 million training units using advanced cloud computing platforms and machine learning algorithms. The database spans the period from 1984 to 2020 and is relevant for research on land cover and land cover change mapping. The database provides high-quality training data and ensures data accuracy through a cross-validation procedure.
Article
Engineering, Multidisciplinary
Hong DanFeng, Wu Xin, Yao Jing, Zhu XiaoXiang
Summary: This study develops a novel superpixel-based subspace learning model that can learn more accurate land cover classification results from multimodal remote sensing data.
SCIENCE CHINA-TECHNOLOGICAL SCIENCES
(2022)
Article
Environmental Sciences
Jianda Cheng, Fan Zhang, Deliang Xiang, Qiang Yin, Yongsheng Zhou, Wei Wang
Summary: This paper introduces a hierarchical capsule network (HCapsNet) for land cover classification of PolSAR images, which considers deep features obtained at different network levels, improving classification performance. By using phase, amplitude, and polarimetric decomposition parameters to uniformly describe scattering mechanisms of different land covers, the generalization performance is enhanced. Additionally, the inclusion of conditional random field (CRF) in the classification framework helps eliminate small isolated regions within classes.
Article
Agronomy
Reija Heinonen, Tuomas J. Mattila
Summary: The study suggests that smartphone-based visual assessment of vegetation cover is a promising method for comparing measurements on different farms. Different smartphones may have varying green reflectance values, but they are correlated with the NDVI and biomass measured by the Sentinel 2 satellite.
Article
Multidisciplinary Sciences
Xinran Li, Daidou Guo, Chuan Qin
Summary: This paper proposes a new cover selection method for steganography, which evaluates the suitability of candidate images for steganography using symmetric steganalytic tools and selects the best candidates. Experimental results show that our method outperforms existing cover selection schemes in steganalytic tool detection.
Article
Engineering, Electrical & Electronic
Nadir Bengana, Janne Heikkila
Summary: Mapping land use and cover is crucial in various fields, with earth observation satellites and deep learning methods playing key roles. Due to the need for large labeled data, domain adaptation is used to address challenges. Significant improvements are achieved through experiments.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2021)
Article
Computer Science, Information Systems
Jintong Jia, Jiarui Song, Qingqiang Kong, Huan Yang, Yunhe Teng, Xuan Song
Summary: This paper proposes a multi-attention-based semantic segmentation network for remote sensing images, addressing the challenges of multiple targets and large feature differences in such images. The model achieves improved extraction capability for fine-grained features by using a coordinate attention-based residual network in the encoder, replaces traditional upsampling operator with a content-aware reorganization module in the decoder to enhance network information extraction, and introduces a fused attention module for feature map fusion to solve the multi-scale problem. Experimental results show superior performance of the proposed model on both WHDLD dataset and self-labeled Lu County dataset, surpassing commonly used benchmark models.
Article
Geochemistry & Geophysics
Gerald Baier, Antonin Deschemps, Michael Schmitt, Naoto Yokoya
Summary: This study uses generative adversarial networks (GANs) to synthesize optical RGB and synthetic aperture radar (SAR) remote sensing images from land cover maps and auxiliary raster data. The research finds that including auxiliary information in the synthesis process improves the quality of the generated images and allows for more control over their characteristics.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Runmin Dong, Weizhen Fang, Haohuan Fu, Lin Gan, Jie Wang, Peng Gong
Summary: This article proposes an online noise correction approach and a synergistic noise correction loss to address the challenges of high-resolution land cover mapping over large areas. By correcting the noisy labels and optimizing the network parameters, the proposed method improves the accuracy of the land cover map from a lower-resolution product in China.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Lianlei Shan, Weiqiang Wang
Summary: Recently, FCN-based networks have achieved impressive success in semantic segmentation of natural images. However, in high-resolution remote sensing image segmentation, there is a considerable gap in accuracy compared to natural images. The key to accurate segmentation is context, and effective networks can obtain large contexts. To address the limitations of networks designed for natural images, targeted improvements including unit fusion and cross-level fusion were proposed. Experimental results on Deepglobe dataset showed significant improvements in segmentation performance.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Environmental Sciences
Elif Sertel, Burak Ekim, Paria Ettehadi Osgouei, M. Erdem Kabadayi
Summary: This research focuses on the deep learning-based segmentation of VHR satellite images, specifically for LULC mapping. The study shows that the DeepLabv3+ architecture with a ResNeXt50 encoder achieves the best performance, providing highly accurate LULC maps.
Article
Environmental Sciences
Xu Shen, Liguo Weng, Min Xia, Haifeng Lin
Summary: This paper proposes a multi-scale feature aggregation network to achieve high-precision land cover segmentation by addressing the issues of traditional semantic segmentation networks.
Article
Remote Sensing
Bingxiao Wu, Zhujun Gu, Wuming Zhang, Qinghua Fu, Maimai Zeng, Aiguang Li
Summary: This study proposes the investigator accuracy (IA) metric for image segmentation validation, focusing on the location accuracy of single patches. It evaluates the capture accuracy of near-center subregions and category weight to determine segmentation quality. Grayscale dilation and erosion algorithms are optimized, and a parallel analysis scheme is applied for efficient IA evaluation. Results show that the capture accuracy and category weight of a patch affect its IA.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2023)
Article
Multidisciplinary Sciences
Jagan Nath Adhikari, Bishnu Prasad Bhattarai, Maan Bahadur Rokaya, Tej Bahadur Thapa
Summary: This study evaluates the land use/land cover dynamics between 2000 to 2020 in the central part of the Chitwan Annapurna Landscape, Nepal using Landsat images. The results show an increase in developed areas, mixed forests, and Sal dominated forests, and a decrease in riverine forests, barren areas, croplands, and grasslands. These findings can be used for wildlife habitat protection and future change forecasting.
Review
Chemistry, Analytical
Shengyu Zhao, Kaiwen Tu, Shutong Ye, Hao Tang, Yaocong Hu, Chao Xie
Summary: LULC image classification is an important component of Earth observation technology. It uses remote sensing techniques to classify ground cover and provides information for environmental protection, urban planning, and land resource management. Deep learning methods have achieved remarkable results in processing remote sensing data, bringing new possibilities for LULC classification research and development.