Project brief
From scenario and baseline to reproducible evidence.
Build a traceable pipeline from LiDAR point-cloud collection and synchronization through labeling, quality control, and a 3D world-model evaluation.
The handbook starts with a runnable minimum, then adds real constraints. The final presentation should explain the method, results, failures, limitations, and next steps.
Evaluation focus
Use metrics to make trade-offs visible.
Project handbook
Project brief
Choose a clear setting such as a campus path, warehouse, indoor corridor, parking area, or small robot workspace. Produce a data card and support at least one downstream task: 3D detection, segmentation, occupancy, scene flow, trajectory prediction, or reconstruction.
Minimum version
Use at least five timestamped sequences or 100 frames, from a public dataset, simulator, or approved static collection.
- A coordinate-system and calibration description.
- A schema for points, intensity, timestamps, poses, and scene labels.
- Visualization for single frames, accumulated clouds, and trajectories.
- Quality checks for empty frames, duplicates, drift, and unusual reflections.
Privacy and safety
Do not collect in private places without permission or publish identifiable people, license plates, addresses, or sensitive areas. Follow campus and venue rules, and attach the dataset license, collection scope, and deletion policy.
Resource note: External code, models, datasets, and platforms remain subject to their own licenses, terms of service, and applicable law.