This project aims to develop a remote sensing-based deep learning approach to systematically map the sand mining activities in large rivers and to quantify, for the first time, the annual sand extraction rates in the Mekong River.
Urbanisation and land reclamation have resulted in sand being stripped from the riverbeds of many major rivers around the world. Although river sand mining is widespread and cause a variety of environmental problems, data on sand mining rates is scant due to difficulties in monitoring sand mining activities. To fill the knowledge gap, this proposal aims to develop a remote sensing-based Deep Learning approach to systematically map the sand mining activities in large rivers and to quantify for the first time, the annual sand extraction rates in the Mekong River. The Mekong River was chosen as a representative case study for the developing our method, because it has been mined intensively in the last 30 years with no signs of abating. We expect that our results will improve our understanding of the environmental impacts arising from sand mining activities, as these impacts are often compounded by the effects of other anthropogenic drivers such as upstream dams and groundwater extraction. Our research will also be useful for determining sustainable mining rates that can be used by local governments to setup regulatory frameworks. Finally, we envision that the novel Deep Learning approach to be developed has high potential to be applied to other large rivers across the world.
Earth Observatory of Singapore
2022 to 2023
Director, Remote Sensing Lab
Ho Huu Loc, Asian Institute of Technology
Feng Lian, Southern University of Science and Technology
Jingyu Wang, National Institute of Education
Charles-Robin Gruel, Centre National de la Recherche Scientifique
Email is required
Email is wrong format