Current methods to map land use in Australia are time and resource intensive. Advances in big data and imagery availability have created a need to develop methods to automatically classify land use features to allow for rapid response, higher spatial and temporal resolution and a more detailed classification. Updated land use mapping and land use change are key spatial data sets and fundamental to the monitoring of land management impacts on water quality, natural disaster recovery and biosecurity planning and response. The objective of this proposed PhD work is to investigate methods using Convolutional Neural Networks (CNN) to automatically classify land use features from high resolution earth observation data.
 

Project members

Andy CLARK

PhD Candidate