|
Getting your Trinity Audio player ready...
|
The interdisciplinary realm of GeoAI, amalgamating geospatial science with artificial intelligence (AI), is at the forefront of addressing pressing environmental and social challenges worldwide. The Hong Kong Polytechnic University (PolyU) has spearheaded pioneering initiatives in this domain, harnessing innovative GeoAI technologies across various sectors such as transportation, urban planning, public safety, climate change, and disaster management.
Under the leadership of Prof. Qihao Weng, Chair Professor of Geomatics and Artificial Intelligence at the Department of Land Surveying and Geo-Informatics, PolyU established the Research Centre for Artificial Intelligence in Geomatics (RCAIG). With a vision to emerge as a global research and development hub in GeoAI, RCAIG focuses on devising original AI methodologies and technologies tailored for geomatics applications in urban areas.
RCAIG’s research thrusts encompass a spectrum of domains ranging from human-environment interactions in urbanisation to the creation of comprehensive data products for global urban areas utilising Earth Observations (EO) and providing EO-based urban data services. Prof. Weng emphasised the pivotal role of Earth observation in comprehending environmental and societal changes, with research spanning geospatial big data, remote sensing, ground-based sensors, navigation, surveying, and photogrammetry.
One pivotal application area where GeoAI has made significant strides is building monitoring. Leveraging thousands of learnable parameters, GeoAI enables automated identification of building characteristics such as colour and shape. This technology finds crucial applications in disaster response, building height estimation, structural change detection, and building energy consumption estimation, revolutionising the efficiency and insightfulness of building monitoring processes.
In environmental monitoring, RCAIG researchers have developed an innovative urban cellular automata (CA) model based on impervious surface area. Utilising annual urban extent time series data from satellite observations, this model simulates fractional changes in urban areas within each grid. By categorising historical urban growth pathways, it offers nuanced insights into urbanisation dynamics, supporting sustainable development efforts with enhanced computational efficiency and performance.
Ms. Wanru He and her team’s research on this model, as reported in the paper “Modeling gridded urban fractional change using the temporal context information in the urban cellular automata model,” published in Cities, showcases its potential for modelling urban growth at regional and global scales under diverse urbanisation scenarios.
In the realm of smart traffic management, RCAIG researchers have devised a multi-agent approach for order matching and vehicle repositioning in ride-hailing services. This innovative technology optimises the coordination of supply and demand, aiming to enhance the overall efficiency of ride-hailing platforms. By employing multi-agent deep reinforcement learning, the approach tackles the complexities of transportation planning, achieving remarkable results such as reduced passenger rejection rates and driver idle time.
Ms. Mingyue Xu and her team’s research, detailed in the paper “Multi-agent reinforcement learning to unify order-matching and vehicle-repositioning in ride-hailing services,” published in the International Journal of Geographical Information Science, demonstrates the transformative potential of this approach in addressing long-term spatiotemporal planning challenges in transportation.
Overall, PolyU’s RCAIG stands at the forefront of GeoAI research, pioneering transformative technologies to tackle multifaceted environmental and social challenges, ranging from urbanisation dynamics to smart transportation management. The centre continues to drive innovation, cementing its position as a global leader in GeoAI research and development.
