On October24, 2025, Zhenghang Yuan successfully defended her doctoral dissertation entitled:
“Advancing Vision-Language Understanding in Remote Sensing with Deep Learning”
The defense took place under the chairmanship of Prof. Dr.-Ing. Muhammad Shahzad (TUM), with the examination committee consisting of:
Prof. Dr.-Ing. habil. Xiaoxiang Zhu (TUM, PhD Supervisor)
Prof. Dr. Sébastien Lefevre (University of South Brittany)
Prof. Dr. Ioannis Papoutsis (National Technical University of Athens, Greece)
About the Dissertation
Zhenghang’s work advances deep learning-based vision-language understanding for EO — exploring how to make complex satellite data more accessible, interpretable, and interactive through language. Her contributions include:
- Language-Guided Curriculum Learning for VQA — teaching models to learn progressively from easy to hard questions, improving robustness and understanding.
- Change Detection Meets VQA — enabling spatiotemporal reasoning to answer questions about environmental and urban dynamics.
- Referring Remote Sensing Image Segmentation (RRSIS) — moving beyond text answers to fine-grained, pixel-level visual explanations.
- ChatEarthNet — a global-scale dataset of 163K+ Sentinel-1/2 image-text pairs, enabling training of vision-language geo-foundation models.
- ChatEarthBench, a new benchmark for fair evaluation of EO VLMs.
Zhenghang’s dissertation embodies originality, and OpenScience, inspiring more research towards a future where humans can interact with satellite data as naturally as they converse with language models.
Congratulations 🎉
We warmly congratulate Dr. Zhenghang Yuan on this outstanding achievement and wish her continued success in advancing the frontiers of LLM and Earth Observation research.