Generation of 3D image point representation of video footage which can be manipulated and moved
The service generates the 3D reconstruction of the structural element being inspected. For this purpose, a set of images of the element are generated from all possible orientations by means of a flight around the element. These images combined with the latitude and longitude data serve as input to the generation algorithm. A monocular version of the SLAM algorithm (Simultaneous localization and mapping) is used to generate the 3D object of the infrastructure.
- Video Splitter: Separates the video into frames at a specific number of frames per second.
- Metadata Decoding: Decodes the UAV's telemetry metadata channel, Mavlink or KLV (key length value) format.
- Geo Annotation: Labels each frame with the latitude, longitude and altitude values of the camera that captured them.
- SLAM 3D: Generates the 3D model from the georeferenced images using SLAM 3D algorithm.
Execution and deployment
- Amazon Web Services
- Microsoft Azure
- Google Cloud
Ground Control Station:
- Atos Bull Sequana