Article
Environmental Sciences
Zeying Lan, Yang Liu, Jianhua He, Xin Hu
Summary: This study addresses the challenges of separating forests and buildings in polarimetric synthetic aperture radar (PolSAR) by introducing new parameters and utilizing the random forest algorithm. The results show significant improvement in classification accuracy through the integration of polarimetric parameters with scattering power.
Article
Geochemistry & Geophysics
Yuwei Guo, Licheng Jiao, Rong Qu, Zhuangzhuang Sun, Shuang Wang, Shuo Wang, Fang Liu
Summary: This paper proposes an adaptive fuzzy superpixel (AFS) algorithm based on polarimetric scattering information for PolSAR image classification. AFS utilizes the correlation between pixels' polarimetric scattering information to generate superpixels, and dynamically updates the ratio of undetermined pixels. Experimental results demonstrate the superiority of AFS in PolSAR image classification problems.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Weixian Tan, Borong Sun, Chenyu Xiao, Pingping Huang, Wei Xu, Wen Yang
Summary: This study introduces an unsupervised classification method based on polarimetric synthetic aperture radar (PolSAR) images, targeted at the particular surface features of the Hunshandake Sandy Land. By utilizing new decomposition and large-scale spectral clustering, the method achieves more accurate feature extraction and complex terrain classification, leading to improved classification performance.
Article
Environmental Sciences
Wenqiang Hua, Yurong Zhang, Cong Zhang, Xiaomin Jin
Summary: Deep learning and convolutional neural networks (CNN) are widely used in PolSAR image classification and have achieved satisfactory results. However, the issue of obtaining labeled samples for PolSAR images remains a challenge due to the time and labor required. To address this, a new attention-based 3D residual relation network (3D-ARRN) is proposed, which extracts depth polarimetric features using a multilayer CNN with a residual structure and improves classification results through a spatial weighted attention network (SWANet). The proposed model achieves higher classification results with fewer labeled data compared to other comparison methods, as demonstrated in studies on four different PolSAR datasets.
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
Environmental Sciences
Yixin Zuo, Jiayi Guo, Yueting Zhang, Bin Lei, Yuxin Hu, Mingzhi Wang
Summary: Convolutional Neural Network models play a vital role in the supervised classification of PolSAR images, but research on unsupervised PolSAR classification is scarce. This paper introduces a completely unsupervised model that combines Convolutional Autoencoder with Vector Quantization, achieving high overall accuracy on satellite and airborne full polarization data.
Article
Computer Science, Artificial Intelligence
Ronghua Shang, Jiaming Wang, Licheng Jiao, Xiaohui Yang, Yangyang Li
Summary: This paper proposes a spatial feature-based convolutional neural network (SF-CNN) for solving PolSAR classification problems. The special structure of SF-CNN can expand the training set by combining different samples and enhance the network's ability to extract discriminative features in low-dimensional feature space. Experimental results show that SF-CNN outperforms standard CNN in PolSAR image classification tasks.
APPLIED SOFT COMPUTING
(2022)
Article
Computer Science, Information Systems
Guofan Shao, Lina Tang, Hao Zhang
Summary: This paper introduces image classification efficacy metrics from the fields of medicine and pharmacology to overcome the limitations of existing accuracy metrics. The method can be applied in various levels and types of image classifications, facilitating accurate evaluation of classification methods.
Article
Geochemistry & Geophysics
Xianyuan Wang, Zongjie Cao, Yiming Pi
Summary: The adaptive anchor graph regularization method proposed in this letter for PolSAR images improves classification accuracy by selecting an optimal number of nearest anchors for each pixel, addressing the limitations of global parameters.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Environmental Sciences
Aili Wang, Shuang Xing, Yan Zhao, Haibin Wu, Yuji Iwahori
Summary: This study proposes a method that combines a spectral spatial kernel and an improved Vision Transformer (ViT) for the classification task of hyperspectral images. By dimensionally reducing and extracting features from the hyperspectral data, and introducing a re-attention mechanism and a local mechanism, the method can better mine and represent the sequence properties of spectral features and utilize the local and global information of the data, thereby improving the classification accuracy.
Article
Computer Science, Information Systems
Amanjot Singh, Jagroop Singh
Summary: This paper presents a new method for image upscaling and de-blocking of compressed images. The proposed technique in the spatial domain is practical and realistic, showing promising results in comparison to other interpolation methods.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Environmental Sciences
Yachao Zhang, Xuan Lai, Yuan Xie, Yanyun Qu, Cuihua Li
Summary: This paper proposes a new discriminative dictionary learning method based on Riemann geometric perception for PolSAR image classification. An optimization model was established, and an efficient optimization algorithm was developed to solve it. Experimental results showed the method's superiority in accuracy and robustness.
Article
Remote Sensing
Wenchun Wu, Yunjie Shi, Kui Cai, Keyu Li, Lhaba Cering, Ziling Gong
Summary: The research utilized an explicit decision tree classification strategy with multiple feature nodes based on Landsat-8 OLI imagery to extract wetland information in QNNR, achieving a high accuracy classification result.
REMOTE SENSING LETTERS
(2021)
Article
Computer Science, Artificial Intelligence
Hiba Alqasir, Damien Muselet, Christophe Ducottet
Summary: The article proposes a solution to improve the learning process of a classification network by providing shape priors, reducing the need for annotated data. The solution is tested on cross-domain digit classification tasks and a video surveillance application.
Article
Geochemistry & Geophysics
Ronghua Shang, Wenjie Li, Weitong Zhang, Jie Feng, Yangyang Li, Licheng Jiao
Summary: This article proposes an adaptive projection attention-based simplified nonlocal neural network for classification of imbalanced samples in hyperspectral images. The network calculates local information and learns global semantic information through a simplified nonlocal network. It also utilizes adaptive projection, multiscale pooling layers, and extended features to alleviate the imbalanced sample issue and improve classification accuracy.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Benyamin Hosseiny, Masoud Mahdianpari, Brian Brisco, Fariba Mohammadimanesh, Bahram Salehi
Summary: This study aims to develop a classification system for mapping complex wetland areas by incorporating deep ensemble learning and satellite datasets. WetNet outperforms state-of-the-art deep models in terms of classification accuracy and processing time.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Ali Jamali, Masoud Mahdianpari
Summary: This study explores the potential and limitations of integrating deep learning networks with transformers for complex coastal wetland classification. The results demonstrate that the multi-model network outperforms other solo classifiers in wetland classification.
Article
Environmental Sciences
Rezvan Habibollahi, Seyd Teymoor Seydi, Mahdi Hasanlou, Masoud Mahdianpari
Summary: This study proposes a deep learning-based change detection algorithm for bi-temporal PolSAR imagery using a transfer learning method. The algorithm can accurately and timely extract changes, and it shows higher accuracy and kappa coefficient compared to other methods in the experiments.
Article
Environmental Sciences
Mohammed Dabboor, Ian Olthof, Masoud Mahdianpari, Fariba Mohammadimanesh, Mohammed Shokr, Brian Brisco, Saeid Homayouni
Summary: The Canadian RADARSAT Constellation Mission (RCM) has completed its early operation phase and is currently undergoing performance evaluation. This study provides an overview of the initial results obtained for three high-priority applications: flood mapping, sea ice analysis, and wetland classification. The study focuses on the performance of the synthetic aperture radar (SAR) using both linear polarization and Compact Polarimetric (CP) architecture. The results show promising agreement between RCM and RADARSAT-2 in flood mapping, improved contrast in sea ice analysis, and accurate wetland classification using fusion of RCM SAR data and optical imagery.
Article
Environmental Sciences
Ali Jamali, Masoud Mahdianpari
Summary: The emergence of deep learning techniques has revolutionized the classification of complicated environments, such as remote sensing. This study employed the Swin Transformer algorithm to classify complex coastal wetlands and compared its performance with the well-known deep CNNs of AlexNet and VGG-16. The results showed that the Swin Transformer algorithm outperformed the other techniques in terms of accuracy and F-1 scores, indicating the high capability of transformers in remote sensing for the classification of complex landscapes.
Article
Environmental Sciences
Haifa Tamiminia, Bahram Salehi, Masoud Mahdianpari, Colin M. Beier, Lucas Johnson, Daniel B. Phoenix, Michael Mahoney
Summary: This study aims to enhance the accuracy of forest above-ground biomass estimation by investigating the performance of remote sensing data sources, examining tree-based machine learning models, and comparing pixel-based and object-based image analysis methods. The results show that combining remote sensing data from multiple sources improves model accuracy, with OBIA providing the best results when combining optical and SAR data.
GEOCARTO INTERNATIONAL
(2022)
Article
Environmental Sciences
Hangyu Lyu, Weimin Huang, Masoud Mahdianpari
Summary: Sea ice monitoring plays a crucial role in navigation safety and offshore activities. Synthetic aperture radar (SAR) has been widely used for sea ice remote sensing due to its ability to collect data in various weather conditions. The RADARSAT Constellation Mission (RCM) is a new Canadian SAR mission that offers improved spatial coverage and temporal resolution. This paper introduces NFNet as a deep convolutional neural network for sea ice detection and classification using RCM data, and demonstrates its superiority over conventional techniques.
Article
Environmental Sciences
Haifa Tamiminia, Bahram Salehi, Masoud Mahdianpari, Colin M. Beier, Lucas Johnson
Summary: In this study, a machine learning-based workflow was proposed to estimate and analyze the historical aboveground biomass (AGB) of forests in New York State using Landsat historical imagery, airborne LiDAR, and forest plot data. The results showed a decrease of 983.79 x 10(6) Mg/ha in deciduous forest AGB from 2001 to 2006, followed by an increase of 618.28 x 10(6) Mg/ha from 2006 to 2011, and a further increase of 229.12 x 10(6) Mg/ha from 2011 to 2016. Finally, there was a slight change in AGB from 2016 to 2019. The proposed framework can be applied to monitoring forests in other states or even on a national scale.
Review
Environmental Sciences
Hamid Jafarzadeh, Masoud Mahdianpari, Eric W. Gill, Brian Brisco, Fariba Mohammadimanesh
Summary: Despite the threats of pollution and development, wetlands play a crucial role in ecosystem services. This paper provides an overview of wetland studies utilizing remote sensing (RS) methods and investigates publications from 1990 up to the middle of 2022. The synthesis of findings suggests that combining RS data and machine learning (ML) algorithms is beneficial for wetland monitoring and research, opening up new perspectives for future studies.
Article
Environmental Sciences
Ali Radman, Masoud Mahdianpari, Brian Brisco, Bahram Salehi, Fariba Mohammadimanesh
Summary: This study proposes a novel deep learning approach that combines convolutional neural networks (CNNs) and graph convolutional networks (GCNs) for PolSAR image classification. The proposed model incorporates spatial-based and polarimetric-based features and achieves improved accuracy compared to conventional methods on the AIRSAR dataset.
Article
Environmental Sciences
Narges Takhtkeshha, Ali Mohammadzadeh, Bahram Salehi
Summary: This paper proposes a novel deep-learning-based method for rapid post-earthquake building damage detection. By combining satellite images obtained by UAVs with automatic selection of training samples and machine learning algorithms, accurate building damage maps can be generated.
Article
Engineering, Electrical & Electronic
Ali Jamali, Masoud Mahdianpari, Fariba Mohammadimanesh, Brian Brisco, Bahram Salehi
Summary: This article introduces a novel approach called 3-D hybrid GAN to address the limited training sample issue in classification, achieving better results in complex wetland classification.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Ali Jamali, Masoud Mahdianpari, Fariba Mohammadimanesh, Avik Bhattacharya, Saeid Homayouni
Summary: This article proposes the use of Haar wavelet transform in deep CNNs for improved classification accuracy of PolSAR imagery. Experimental results show that the proposed method outperforms other shallow CNN models in terms of accuracy and consistency of classification results.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
