Qidi Shu

Qidi Shu

The Hong Kong Polytechnic University

Hi there! I am Qidi Shu (舒奇迪), a PhD candidate from Hong Kong PolyU.

I am interested in AI for a Dynamic Earth, with a focus on understanding, reconstructing, and reasoning about Earth surface changes from remote sensing images across time.

My research spans remote sensing image interpretation, multimodal data fusion, vision foundation models, and lies at the intersection of discriminative understanding and generative modeling.

I am particularly interested in bi-temporal and timeseries remote sensing, with the broader goal of moving beyond static perception toward modeling and understanding a changing Earth.

（I am currently seeking opportunities in both academia and industry. Please feel free to reach out for potential collaborations or opportunities 🤝🤝 Email）

🔥 News

  • 2026.10:   Our work RESTORE-DiT is marked as highly cited paper in web of science!
  • 2026.09:   One co-author work of 30-year China Settlement Footprint mapping is accepted by Scientific Data! The dataset is available at Link.
  • 2026.02:   I have completed my six-month visit at SNU Ecological Sensing AI Lab. A fruitful and happy journey! Thank you, Prof. Ryu! I will miss all of you guys.
  • 2025.08:   Attended IGARSS 2025 in Brisbane, Australia! Shake Hands with a Kangaroo 🦘
  • 2025.06:   Our work RESTORE-DiT has been accepted by Remote Sensing of Environment. We propose a diffusion model-based optical-SAR fusion framework for Sentinel-2 time series image reconstruction.

📝 Publications

RSE 2025
sym

RESTORE-DiT: Reliable satellite image time series reconstruction by multimodal sequential diffusion transformer

Qidi Shu, Xiaolin Zhu, Shuai Xu, Yan Wang, Denghong Liu

Code

  • RESTORE-DiT is a novel Diffusion-based framework for Satellite Image Time Series (SITS) reconstruction. Our work firstly promotes the sequence-level optical-SAR fusion through a diffusion framework.
TGRS 2024
sym

CutMix-CD: Advancing Semi-Supervised Change Detection via Mixed Sample Consistency

Qidi Shu, Xiaolin Zhu, Luoma Wan, Shuheng Zhao, Denghong Liu, Longkang Peng, Xiaobei Chen

Code

  • To address the issue of high labeling costs in bi-temporal image change detection tasks, we propose a novel semi-supervised method CutMix-CD, which incorporates the pair-level change-aware CutMix augmentation into the consistency learning framework.
JAG 2022
sym

MTCNet: Multitask consistency network with single temporal supervision for semi-supervised building change detection

Qidi Shu, Jun Pan, Zhuoer Zhang, Mi Wang

Code

  • Single-temporal labels serve as a low-cost alternative when change labels are limited! We present a new task called “Semi-supervised change detection with single temporal supervision”. We propose a multitask consistency network (MTCNet), which fully takes the advantage of easily obtained single-temporal labels for bi-temporal change detection.
JAG 2022
sym

DPCC-Net: Dual-perspective change contextual network for change detection in high-resolution remote sensing

Qidi Shu, Jun Pan, Zhuoer Zhang, Mi Wang

Code

  • DPCCNet emphasizes the process of extraction and optimization of change features by bi-temporal feature fusion and contextual modeling. The dual-perspective fusion (DPF) takes bi-temporal features as reference respectively to increase the sensitivity to change information. The change context module (CCM) incorporates abundant contexts to facilitate the integrity of change objects.

🎖 Awards

  • PolyU Research Excellence Scholarship, candidate of HKPFS, 2023.
  • The 2nd Prize (Rank: 2/181 teams) in the International High Resolution Remote Sensing Image Interpretation Competition. 2023.
  • Outstanding graduate of Wuhan University (Master), 2023.
  • Academic scholarship (First class), Wuhan University, both in 2021 and 2022.
  • Outstanding graduate of UESTC (Bachelor), 2020.
  • Outstanding Student Scholarship, UESTC, both in 2017 and 2018.

📖 Educations

  • 2025.09 - 2026.02, Visiting PhD in SNU (Seoul National University) Ecological Sensing AI Lab, supervised by Prof. Youngryel Ryu.
  • 2023.09 - now, Pursing PhD degree in PolyU (The Hong Kong Polytechnic University), PRIDE lab, supervised by Prof. Xiaolin Zhu.
  • 2020.09 - 2023.06, Master’s degree in WHU (Wuhan University), research on computer vision and remote sensing intelligent interpretation, supervised by Prof. Jun Pan and Prof. Mi Wang.
  • 2016.09 - 2020.06, Bachelor’s degree in UESTC (University of Electronic Science and Technology of China), major in Spatial Information and Digital Technology, supervised by Prof. Xingwen Quan and Prof. Binbin He.

💬 Review experience

  • Reviewer for remote sensing and image processing journals, including Remote Sensing of Environment, IEEE Transactions on Geoscience and Remote Sensing, International Journal of Applied Earth Observation and Geoinformation, Journal of Remote Sensing, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, IEEE Geoscience and Remote Sensing Letters, Scientific Report, Knowledge-based Systems, Neurocomputing, etc.