IGARSS 2022

283 papers
The 2022 edition of the IEEE International Geoscience and Remote Sensing Symposium (IGARSS) featured 315 papers addressing diverse challenges in Earth observation and data analysis. Research themes focused heavily on the application of deep learning architectures, including transformers and contrastive learning, for hyperspectral image classification and point cloud downsampling. Methodological trends included the development of heterogeneous image matching algorithms for UAV and satellite data, as well as pan-sharpening techniques utilizing wavelet-assisted fusion models. Applications in environmental monitoring and object detection were also prominent. Studies explored snow water equivalent estimation through differential SAR interferometry and the use of conditional adversarial domain adaptation for forest inventory classification. Additionally, the proceedings highlighted advancements in detecting small vehicles and ships within SAR and optical imagery. Prolific contributors to this year's research included Yanfei Zhong, Licheng Jiao, and Liangpei Zhang. The collective work reflects a shift toward integrating complex spatial-spectral sampling and adaptive embedding techniques to improve the accuracy of remote sensing interpretations.

Top Research Directions

Research directions most represented by IGARSS 2022 papers.

Cross Dataset Validation for Robust Model Generalization 2
Inferring Surface Emissions Through Atmospheric Inversion 1
Mitigating Radar Layover and Topographic Illumination Artifacts 1
Neural Similarity Learning with Trainable Metrics 1
Supervised Dimensionality Reduction 1
Replacing Fully Connected Layers with Global Average Pooling 1
Global Covariance Pooling for Deep Feature Aggregation 1
Robust Information Fusion for Highly Conflicting Evidence 1
Overcoming Pseudo Changes in Remote Sensing Change Detection 1
Spectral Signal Conditioning for Robust Quantitative Analysis 1
Mechanisms for Rapid Iterative Convergence 1
Resolution Agnostic Spatial Encoding for Vision Transformers 1
Joint Optimization of Feature Representations and Binary Hashing 1
Bridging the Spatial Scale Mismatch in Climate Projections 1
Hybrid Uncertainty Quantification in Deep Learning 1

Top Authors

AuthorPapers
Licheng Jiao9
Yanfei Zhong9
Xu Tang8
Jocelyn Chanussot7
Jia Liu5
Jonathan Li5
Liangpei Zhang5
Qiangqiang Yuan5
Shaohui Mei5
Yuxin Sun5
Cheng Wang4
Dino Ienco4
Fang Liu4
Jean-michel Morel4
Man On Pun4
Xiangrong Zhang4
Yang Xu4
Yilei Shi4
Zhihui Wei4
Biao Hou3

Impactful Papers

Papers from IGARSS 2022 that shaped research clusters.

Foundational

Influential

Papers

Title
Embedding Ethics and Trustworthiness for Sustainable AI in Earth Sciences: Where Do We Begin?
Enhanced Single-Shot Detector for Small Object Detection in Remote Sensing Images
Enhance Essential Features for Road Extraction from Remote Sensing Images
Estimating Forest Heights and Wood Volume using a Deep Learning Approach from Gedi Waveform Data
Estimation of Atmospheric PM2.5 based on Photos and Deep Learning
Estimation of Hourly Air Temperature in China Based on LightGBM and Himawari-8
Estimation of PM2.5 and PM10 Mass Concentrations in Mining City Cluster from Gaofen-L Aerosol Optical Depth data and Chemical Transport Model
Estimation of Surface-Level Ozone Mass Concentration Using Tropomi Data and Source-Sink Analysis Over China
Evaluation of Urban Functional System
Evolutionary Tensor Train Decomposition for Hyper-Spectral Remote Sensing Images
Explainable Analysis of Deep Learning Methods for Sar Image Classification
Exploration of Data Acquisition and Processing Technology Course Education Based on OBE Concept
Exploration on the cultivation of innovative talents in electronic circuit
Exploring Similarity in Polarization: Contrastive Learning with Siamese Networks for Ship Classification in Sentinel-1 SAR Images
Feature and Output Consistency Training for Semi-Supervised Building Footprint Generation
Feature Integration-Based Training for Cross-Domain Hyperspectral Image Classification
Feedback Convolution Based Autoencoder for Dimensionality Reduction in Hyperspectral Images
Fe-LinkNet: Enhanced D-LinkNet with Attention and Dense Connection for Road Extraction in High-Resolution Remote Sensing Images
Few-Shot Hyperspectral Image Classification Based on Domain Adaptation of Class Balance
Focusing Azimuth Periodically Interrupted SAR Echo with Deconvolution by Complex FISTA
Foreground-Background Segmentation of Sequential Point Clouds
Gaussian Information Entropy based band Reduction for Unsupervised Hyperspectral Video Tracking
Generating Adversarial Examples on Sar Images by Optimizing Flow Field Directly in Frequency Domain
Generation of Multi-Source Satellite SST Fusion Product of Arctic with High Temporal Resolution
Generation of Nonlinear Internal Waves Around Hainan Island
Generative Adversarial Network for SAR-to-Optical Image Translation with Feature Cross-Fusion Inference
Global and Multi-Scale Feature Learning for Remote Sensing Scene Classification
Graph Laplacian Regularized Spectral-Spatial-Sparse Unmixing for Hyperspectral Imagery
GRE and Beyond: A Global Road Extraction Dataset
Guiding Field Exploration on Earth and Mars with Outlier Detection
Heterogeneous Image Classification with Multi-Stage Conditional Adversarial Domain Adaptation Between SAR and Optical Imagery
HFGAN: A Heterogeneous Fusion Generative Adversarial Network for Sar-to-Optical Image Translation
HFL-Net: Highlight Foreground and Local Scale Features Network for Cross-Domain Ship Detection
High-Resolution Insar Imaging Via Cs-Based Amplitude-Phase Separation Algorithm
High Resolution Remote Sensing Image Semantic Segmentation Based on Ultra-Lightweight Fully Convolution Neural Network
High Resolution Satellite Sar Focusing with Azimuth Periodically Gapped RAW Data
HTD-VIT: Spectral-Spatial Joint Hyperspectral Target Detection with Vision Transformer
Humans are Poor Few-Shot Classifiers for Sentinel-2 Land Cover
Hybrid Vision Transformer Model for Hyperspectral Image Classification
Hyperspectral and LiDAR Data Land-Use Classification Using Parallel Transformers
Hyperspectral and Multispectral Data Fusion with 1D-Convolution on Spectrum
Hyperspectral Classification with Gradient Based Active Learning
Hyperspectral Image Classification Using Hierarchical Spatial-Spectral Transformer
Hyperspectral Image Denoising Based on Parallel Cross-Fusion Network
Hyperspectral Image Generation From Rgb Images With Semantic and Spatial Distribution Consistency
Hyperspectral Image Restoration Based on Tensor Average Rank and Spatial-Spectral Total Variation
Hyperspectral Mixed Noise Reduction Using Two-Stage Cascade Refined Network
Hyperspectral Target Detection Based on Tensor RPCA with Constrained Energy Regularization
Hyperspectral Target Detection via Ensemble Learning Deep Multiple Instance Neural Network
Identifying a Composition-Related TB Anomlay in Copernicus Crater Using CE-2 MRM Data

« prev Page 3/6 next »