IGARSS 2026 || Washington, D.C. || 9

THP2.PA: Deep Learning for Remote Sensing Image Analysis and Understanding

Poster

Thu, 13 Aug, 15:00 - 16:15

Location: TBD

Session Type: Poster

Session Co-Chairs: Francesca Bovolo, Fondazione Bruno Kessler and Haonan Guo,

Track: AI and Big Data

THP2.PA.1: SEMANTIC SEGMENTATION OF SEABED FORMS IN THE BALTIC SEA USING PYRAMID VISION TRANSFORMER AND BATHYMETRIC DATA

Paweł Nadachowski, Zbigniew Łubniewski, Gdańsk University of Technology, Poland; Radosław Wróblewski, Gdynia Maritime University, Poland; Karolina Trzcińska, Jarosław Tęgowski, University of Gdańsk, Poland

THP2.PA.2: AUTOMATED SEGMENTATION OF LARGE-SCALE PHOTOVOLTAIC FARMS USING SENTINEL-2 SWIR BANDS AND DEEP LEARNING U-NET++ ARCHITECTURE

Jakub Tobolski, Mateusz Skała, Jakub Świegoda, Miłosz Reterski, Mieszko Warjas, Gdansk University of Technology, Poland

THP2.PA.3: A U-NET DENOISING APPROACH FOR DEPOLARIZATION RATIO ESTIMATES

Amanda Cresanti, University of Maryland, United States; John Yorks, NASA Goddard Space Flight Center, United States; Zhanqing Li, University of Maryland, United States

THP2.PA.4: Manifold-Constrained Residual Learning for Remote Sensing Image Deblurring

Wenyi Zhang, Xiyu Wang, Xiaoyi Yin, Qingwei Chu, Kai Wang, Jianyi Ren, The Aerospace Information Research Institute (AIR) under the Chinese Academy of Sciences (CAS), China

THP2.PA.5: RESEARCH ON THE METHOD OF EXTRACTING INDIVIDUAL BUILDINGS FROM OBLIQUE PHOTOGRAMMETRIC POINT CLOUDS ASSISTED BY DIGITAL ORTHOPHOTO MAPS

Chaoyang Bian, Fang Huang, Qingzhe Lv, Tongtong Lu, Hao Guan, University of Electronic Science and Technology of China, China; Jian Wang, Aerospace Information Research Institute, Chinese Academy of Sciences, China

THP2.PA.6: UNSUPERVISED AUGMENTATION FOR IMPROVING SEGMENTATION ROBUSTNESS ON URBAN REMOTE SENSING DATA

Rohan Reddy, Fraunhofer USA, United States; Kevin Qiu, Fraunhofer IOSB, United States; Arjun Srinivasan, Madeline Diep, Fraunhofer USA, United States; Dimitri Bulatov, Fraunhofer IOSB, Germany

THP2.PA.7: RESEARCH ON SINGLE-TREE SEGMENTATION OF URBAN 3D POINT CLOUDS BASED ON DIRECTION EMBEDDING NETWORK

Zhiming Liu, Fang Huang, Tongtong Lu, Hao Guan, Chaoyang Bian, University of Electronic Science and Technology of China, China; Jian Wang, Chinese Academy of Sciences, China

THP2.PA.8: Oriented Vehicle Detection in SAR Imagery by Leveraging Rotational Convolution and KLD Optimization

Dan Gao, Xiaofang Wu, Zhijin Wen, Academy of Military Sciences of the People's Liberation Army, China

THP2.PA.9: STAGE-DEPENDENT FREQUENCY-DOMAIN REFINEMENT FOR SAR OBJECT DETECTION

Juhyeon Han, Sunok Kim, Korea Aerospace University, Korea (South)

THP2.PA.10: Fine-Grained Pixel-Wise Hyperspectral Land-Cover Classification in Large-Scale Scenes Using Patch-Based Hypergraph Feature Enrichment

Muhammad Husnain, Griffith University, Australia; Ali Zia, La Trobe University, Australia; Vivien Rolland, CSIRO, Australia; Jun Zhou, Griffith University, Australia

THP2.PA.11: Convolutional Autoencoders for Landsat Single-Time Spatial-Spectral Data Compression: Implications for Classification Accuracy

Ahmadreza Safaeinia, Giorgos Mountrakis, State University of New York College of Environmental Science and Forestry, United States

THP2.PA.12: UNSUPERVISED HYPERSPECTRAL UNMIXING VIA AUTOENCODER-GUIDED ELLIPTIC GRAPH ATTENTION NETWORK

Roshan Bhamare, Arun P V, Indian Institute of Information Technology Sri City, India