---
pretty_name: ScanBot
license: cc-by-4.0
language:
  - en
tags:
  - robotics
  - instruction-following
  - vision-language
  - laser-scanning
  - multimodal
task_categories:
  - robotics
dataset_info:
  features:
    - name: file_name
      dtype: string
    - name: rgb_gopro
      dtype: string
    - name: instruction
      dtype: string
    - name: tcp_pose
      dtype: string
    - name: joint_states
      dtype: string
data_files:
  - path: train/train.jsonl
---



# ScanBot: Towards Intelligent Surface Scanning in Embodied Robotic Systems

## 🧠 Dataset Summary
**ScanBot** is a large-scale, multimodal dataset for studying instruction-driven surface scanning in industrial settings. Collected with a UR3 robotic arm and a KEYENCE LJ-X8200 laser profiler, ScanBot captures RGB-D images, laser point clouds, and 6-DOF robot trajectories for a variety of object features and scanning tasks.


## 📦 Use Cases

- Vision-Language Action Planning (VLA)
- Instruction-Guided Surface Scanning
- 3D Surface Reconstruction
- Spatial Reasoning and Feature Localization
- Laser Profile Analysis for Inspection Tasks



## 🗂️ Data Description
```
scanbot/
├── cube1/
│   ├── top_surface/
│   │   ├── path_001/
│   │   │   ├── rgb/
│   │   │   │   └── ...
│   │   │   ├── depth/
│   │   │   │   └── ...
|   |   |   |── 1746226187.997976_gopro.mp4  
│   │   │   ├── robot_joint_states.csv
│   │   │   ├── robot_tcp_poses.csv
│   │   │   └── metadata.json
│   │   ├── path_002/
│   │   │   └── ...
├── cube2
├── cylinder_red
├── cylinder_white
├── cylinder_white
```

## 💬 Task Levels

Each scan is driven by an instruction that falls under one of the following task categories:

| Task Type | Description |
|-----------|-------------|
| T1        | Surface Scan |
| T2        | Geometry Focus |
| T3        | Spatial Reference |
| T4        | Functional Target |
| T5        | Defect Inspection |
| T6        | Comparative Analysis |

## 🛠️ Hardware Setup

<img src="Figures/setup.png" alt="ScanBot Example" width="30%"/>