Update README.md
Browse files
README.md
CHANGED
|
@@ -1,74 +1,132 @@
|
|
| 1 |
---
|
| 2 |
license: mit
|
| 3 |
pretty_name: SynLiDAR
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
---
|
|
|
|
| 5 |
# SynLiDAR Dataset
|
| 6 |
|
| 7 |
## Overview
|
| 8 |
|
| 9 |
-
SynLiDAR is a synthetic LiDAR dataset designed for autonomous driving research.
|
| 10 |
It contains high-quality simulated point cloud sequences and corresponding semantic annotations.
|
| 11 |
|
| 12 |
The dataset provides two variants:
|
| 13 |
-
|
| 14 |
-
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
|
| 16 |
## Dataset Structure
|
|
|
|
| 17 |
```
|
| 18 |
-
SynLiDAR/
|
| 19 |
-
├── FullDataset/
|
| 20 |
-
│ ├── sequences/
|
| 21 |
-
│ │ ├── 00.zip
|
| 22 |
-
│ │ ├──
|
| 23 |
-
│ │ └── 12.zip
|
| 24 |
-
│ └── readme.txt
|
| 25 |
-
│
|
| 26 |
-
├── SubDataset/
|
| 27 |
-
│ ├── sequences/
|
| 28 |
-
│ │ ├── 00.zip
|
| 29 |
-
│ │ ├──
|
| 30 |
-
│ │ └── 12.zip
|
| 31 |
-
│ └── readme.txt
|
| 32 |
-
│
|
| 33 |
-
├── annotations.yaml
|
| 34 |
-
├── read_data.py
|
| 35 |
-
└── README.md
|
| 36 |
```
|
| 37 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 38 |
## Contents
|
| 39 |
-
- sequences/*.zip — Each zip contains LiDAR frames
|
| 40 |
-
- annotations.yaml — Semantic categories and their mappings
|
| 41 |
-
- read_data.py — Example Python loader
|
| 42 |
-
- readme.txt — Additional notes from the original dataset
|
| 43 |
|
| 44 |
-
|
| 45 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 46 |
```
|
| 47 |
-
|
| 48 |
-
|
| 49 |
-
|
| 50 |
-
|
| 51 |
-
|
| 52 |
-
|
| 53 |
-
|
| 54 |
-
|
| 55 |
-
|
| 56 |
-
|
| 57 |
-
|
| 58 |
-
|
| 59 |
-
|
| 60 |
-
|
| 61 |
-
|
| 62 |
-
|
| 63 |
-
|
| 64 |
-
|
| 65 |
-
|
| 66 |
-
|
| 67 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 68 |
```
|
| 69 |
|
| 70 |
-
⚠️ **Disclaimer**: This dataset is intended for research and educational usage.
|
| 71 |
-
Make sure to respect local regulations when training or deploying autonomous driving systems.
|
| 72 |
|
| 73 |
## Citation
|
| 74 |
```
|
|
@@ -81,4 +139,8 @@ Make sure to respect local regulations when training or deploying autonomous dri
|
|
| 81 |
pages={2795--2803},
|
| 82 |
year={2022}
|
| 83 |
}
|
| 84 |
-
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: mit
|
| 3 |
pretty_name: SynLiDAR
|
| 4 |
+
tags:
|
| 5 |
+
- lidar
|
| 6 |
+
- synthetic
|
| 7 |
+
- point-cloud
|
| 8 |
+
- autonomous-driving
|
| 9 |
+
- semantic-segmentation
|
| 10 |
+
- 3D-segmentation
|
| 11 |
---
|
| 12 |
+
|
| 13 |
# SynLiDAR Dataset
|
| 14 |
|
| 15 |
## Overview
|
| 16 |
|
| 17 |
+
**SynLiDAR** is a synthetic LiDAR dataset designed for autonomous driving research.
|
| 18 |
It contains high-quality simulated point cloud sequences and corresponding semantic annotations.
|
| 19 |
|
| 20 |
The dataset provides two variants:
|
| 21 |
+
|
| 22 |
+
- **FullDataset** — complete version for large-scale experiments (branch: `full`)
|
| 23 |
+
- **SubDataset** — smaller version suitable for prototyping, debugging, and benchmarking (branch: `sub`)
|
| 24 |
+
|
| 25 |
+
This HuggingFace repository uses **branches** to separate large data files and metadata:
|
| 26 |
+
|
| 27 |
+
- `main` → metadata, scripts, annotations
|
| 28 |
+
- `full` → FullDataset
|
| 29 |
+
- `sub` → SubDataset
|
| 30 |
+
|
| 31 |
+
---
|
| 32 |
|
| 33 |
## Dataset Structure
|
| 34 |
+
|
| 35 |
```
|
| 36 |
+
SynLiDAR/
|
| 37 |
+
├── FullDataset/
|
| 38 |
+
│ ├── sequences/
|
| 39 |
+
│ │ ├── 00.zip
|
| 40 |
+
│ │ ├── …
|
| 41 |
+
│ │ └── 12.zip
|
| 42 |
+
│ └── readme.txt
|
| 43 |
+
│
|
| 44 |
+
├── SubDataset/
|
| 45 |
+
│ ├── sequences/
|
| 46 |
+
│ │ ├── 00.zip
|
| 47 |
+
│ │ ├── …
|
| 48 |
+
│ │ └── 12.zip
|
| 49 |
+
│ └── readme.txt
|
| 50 |
+
│
|
| 51 |
+
├── annotations.yaml
|
| 52 |
+
├── read_data.py
|
| 53 |
+
└── README.md
|
| 54 |
```
|
| 55 |
|
| 56 |
+
> ⚠️ Some large sequences have been split into chunks (e.g. `06_part01.zip`, `06_part02.zip`, …) to avoid exceeding file size limits.
|
| 57 |
+
> These parts are functionally equivalent to the original `06.zip`.
|
| 58 |
+
|
| 59 |
+
---
|
| 60 |
+
|
| 61 |
## Contents
|
|
|
|
|
|
|
|
|
|
|
|
|
| 62 |
|
| 63 |
+
- `sequences/*.zip` — Each zip contains LiDAR frames from a single drive sequence
|
| 64 |
+
- `annotations.yaml` — Semantic categories and label mappings
|
| 65 |
+
- `read_data.py` — Example Python loader to read `.bin` point cloud files
|
| 66 |
+
- `readme.txt` — Original dataset notes
|
| 67 |
+
|
| 68 |
+
---
|
| 69 |
+
|
| 70 |
+
# 🔽 How to Download the Dataset
|
| 71 |
+
|
| 72 |
+
## Install huggingface_hub
|
| 73 |
+
|
| 74 |
+
```bash
|
| 75 |
+
pip install huggingface_hub
|
| 76 |
```
|
| 77 |
+
## Download Full Dataset (branch: full)
|
| 78 |
+
|
| 79 |
+
```python
|
| 80 |
+
from huggingface_hub import snapshot_download
|
| 81 |
+
|
| 82 |
+
path = snapshot_download(
|
| 83 |
+
repo_id="AR-X/SynLiDAR",
|
| 84 |
+
revision="full",
|
| 85 |
+
)
|
| 86 |
+
print(path)
|
| 87 |
+
```
|
| 88 |
+
|
| 89 |
+
## Download Sub Dataset (branch: sub)
|
| 90 |
+
|
| 91 |
+
```python
|
| 92 |
+
from huggingface_hub import snapshot_download
|
| 93 |
+
|
| 94 |
+
path = snapshot_download(
|
| 95 |
+
repo_id="AR-X/SynLiDAR",
|
| 96 |
+
revision="sub",
|
| 97 |
+
)
|
| 98 |
+
print(path)
|
| 99 |
+
```
|
| 100 |
+
## 🔧 Merging Split ZIP Files
|
| 101 |
+
|
| 102 |
+
Some sequences in the full branch were too large (>50GB) and are stored as multiple parts:
|
| 103 |
+
|
| 104 |
+
Example:
|
| 105 |
+
```
|
| 106 |
+
06_part01.zip
|
| 107 |
+
06_part02.zip
|
| 108 |
+
...
|
| 109 |
+
06_part11.zip
|
| 110 |
+
```
|
| 111 |
+
|
| 112 |
+
These parts correspond to a single original archive. You can merge them into a single folder using Python:
|
| 113 |
+
|
| 114 |
+
```python
|
| 115 |
+
import zipfile
|
| 116 |
+
from pathlib import Path
|
| 117 |
+
|
| 118 |
+
sequence_dir = Path("SynLiDAR/FullDataset/sequences")
|
| 119 |
+
parts = sorted(sequence_dir.glob("06_part*.zip"))
|
| 120 |
+
|
| 121 |
+
output = Path("merged_06")
|
| 122 |
+
output.mkdir(exist_ok=True)
|
| 123 |
+
|
| 124 |
+
for z in parts:
|
| 125 |
+
print("Extracting", z.name)
|
| 126 |
+
with zipfile.ZipFile(z, "r") as zp:
|
| 127 |
+
zp.extractall(output)
|
| 128 |
```
|
| 129 |
|
|
|
|
|
|
|
| 130 |
|
| 131 |
## Citation
|
| 132 |
```
|
|
|
|
| 139 |
pages={2795--2803},
|
| 140 |
year={2022}
|
| 141 |
}
|
| 142 |
+
```
|
| 143 |
+
|
| 144 |
+
|
| 145 |
+
⚠️ **Disclaimer**: This dataset is intended for research and educational usage.
|
| 146 |
+
Make sure to respect local regulations when training or deploying autonomous driving systems.
|