Autonomous trucking introduces dynamic inter-rig geometry due to tractor–trailer articulation, breaking the static calibration assumption used in most multi-view perception systems.
dCAP is a vision-based framework for online 6-DoF tractor–trailer pose estimation and articulated perception.
The official code, pretrained models, and the STT4AT benchmark will be released soon.
- 2026/02/21 🎉 Paper accepted to CVPR 2026
- Code and dataset release coming soon in February, 2026