Background and Current Practice
Photogrammetry is a workflow to convert photographs or videos of physical space into a virtual 2D or 3D model that can be measured and manipulated. Drone photogrammetry is already being conducted on a weekly basis for site monitoring.
The process to produce drone photogrammetry is time consuming and tedious. It takes several days of flying the drone to capture enough input photos.
The flight is semi-automated and still requires a pilot to manage the flight schedule, battery life, and any potential risk to the drone, such as weather.
The data processing currently requires a person to download the data from the drone manually and upload it to the photogrammetry software for processing. Any technical issues are hard to resolve due to the lack of local support.
Opportunities and Key Challenges
Drone photogrammetry is already being done, but the current processes could be further improved. We are interested in a solution that can enhance the drone flight operations and reduce the need for human intervention by intelligently managing the flight schedule, battery life, and risks. It is also desirable if the total flight time can be reduced while maintaining high-quality photogrammetry results. If the data processing can be optimised or automated, the photogrammetry results can be produced more seamlessly and relieve the personnel of the routine activity.
The photogrammetry results would allow site personnel to visualise topographical surveys of the site and compare the results based on the different timestamps. 3D visualisation would be useful to study some key indicators of progress, such as the excavation depth and building height. A web application to display the photogrammetry results is preferred.
The challenge statement owner’s own drones are available for use, but the solution provider could propose its own drone if it supports better integration with the proposed solution.
The solution replaces the existing method to produce drone photogrammetry by enhancing the drone flight operation and data processing, while minimising the need for human intervention.