Lixian Zhang

Lixian Zhang ｜ 张立贤

PhD candidate in
Department of Earth System Science
Tsinghua University

Email: zhanglixian18@mails.tsinghua.edu.cn

Education Background

Sep 2014 – June 2018
Wuhan University, B.S. in School of Geodesy and Geomatics

Sep 2018 – Ongoing
Tsinghua University, Ph.D. candidate

Intern Experience

May 2020 – Sep 2020
TerraQuanta, Computer vision algorithm intern

Sep 2020 – Ongoing
SenseTime, Remote sensing algorithm intern

Research Interest

Infrastructure Extraction based on Remote Sensing and Deep Learning

Road extraction, Building extraction, Street trees detection

• We propose a promising sub-branch learning to improve extraction accuracy in two stages: 1) the multi-task learning stage and 2) the post-process stage for iterative road connectivity refinement
  
• We propose a novel framework that integrates super-resolution and image segmentation for achieving super-resolution building extraction
  
• We integrate data of overhead view (remote sensing images) and eye-level view (Google Street View) to detect urban trees using Faster R-CNN and saliency detection
  

Remote Sensing Images Quality Improvement

Super resolution, Image deblurring, Perceptional improvement of remote sensing

• We propose a solid reference-based super-resolution approapch for remote sensing image. (Dong et al., 2020 IEEE TGRS, accept)
• We propose a potential approach for cross-sensors super resolution for medium-resolution satellite images based on deep learning.

Competitions & Awards

Huawei Cloud Competition: forecasting short-term rainfall using massive radar echo data, Third Prize, 2020
Tsinghua University Secondary Scholarship, 2020
National Scholarship, 2018
National Scholarship, 2018