SLAM 3D map
Workload class: 3D mapping and SLAM from sensor submissions — mesh fidelity, semantic structure, and agreement with ground truth where the subnet defines it.
Miners compete on reconstruction quality. Sensor bundles (camera, IMU, LiDAR, or other channels as required) are part of the PoPW record; validators score outputs independently of the operator who captured the data.
Testnet
The public dashboard URL for this workload will be announced with the testnet release.
Task shape (illustrative)
Tasks may request maps, reconstructions, semantic labeling, or comparison to held-out references, depending on subnet configuration. Use the live dashboard for the current fields and testKNX costs when URLs are published.
Why this matters
3D and SLAM outputs support mine engineering, built-environment reviews, and industrial digital twins. Verifiable provenance matters: not only a convincing mesh, but evidence that data came from a real run aligned to a defined instruction.
See also
ORB-SLAM3 & RTAB-Map — reference SLAM / 3D mapping stacks
Last updated