Bio:

Dr. Chun Liu is a Professor with the College of Surveying and Geo-Informatics, Tongji University, China. He serves as the Executive Vice Dean of National College of Elite Engineers and the Vice Dean of Graduate School, Tongji University. His research interests include low-altitude ecological sensing and multi-dimensional modeling using remote sensing and AI techniques,with a focus on urban complex environments.His research has been recognized widely and won several prestigious awards by Chinese government.

Talk Title:

Intelligent Perception and Scene Cognition of Disaster Surveying with Quadruped Robotics

Abstract:

China frequently experiences geological disasters. There are currently nearly 300,000 identified geological disaster hazard sites nationwide, threatening about 20 million people and causing property losses of roughly 450 billion yuan. Consequently, current research focuses on establishing 'Space-Air-Ground' landslide monitoring networks. However, existing technologies face blind spots in landslide disaster perception: satellite remote sensing data often fall short in resolution and accuracy for detailed terrain characterization; UAVs struggle to access sites due to spatial obstacles and limited endurance; and manual investigations entail safety risks. These factors result in a lack of real-time core disaster area data, severely restricting landslide mechanism analysis and precursor identification. To address these limitations, this presentation centers on quadrupedal measurement robots to advance perception capabilities in typical landslide scenarios. We developed a tripartite research framework encompassing equipment, algorithms, and applications, with investigations focusing on: (i) observation task planning, (ii) multi-source observation fusion, and (iii) semantic hazard identification. This aims to bridge the data perception gap within the 'last mile' of landslide disaster zones.