File size: 3,836 Bytes
d332fcc
 
 
a903b39
 
 
 
 
 
 
d332fcc
a903b39
45d9210
 
 
 
a903b39
45d9210
 
 
a903b39
 
 
 
 
 
 
 
 
 
 
45d9210
 
a903b39
45d9210
a903b39
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
45d9210
 
a903b39
 
 
 
 
45d9210
 
a903b39
 
 
 
 
 
 
 
 
 
 
 
 
45d9210
a903b39
 
 
 
 
 
 
98f766c
a903b39
98f766c
a903b39
 
 
 
 
 
 
 
 
 
 
98f766c
a903b39
98f766c
a903b39
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
036547d
 
a903b39
036547d
 
 
a903b39
036547d
 
 
 
 
 
 
45d9210
 
 
 
 
 
 
 
 
 
 
 
 
 
a903b39
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
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
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
---
license: mit
pretty_name: SynLiDAR
tags:
  - lidar
  - synthetic
  - point-cloud
  - autonomous-driving
  - semantic-segmentation
  - 3D-segmentation
---

# SynLiDAR Dataset

## Overview

**SynLiDAR** is a synthetic LiDAR dataset designed for autonomous driving research.  
It contains high-quality simulated point cloud sequences and corresponding semantic annotations.

The dataset provides two variants:

- **FullDataset** — complete version for large-scale experiments (branch: `full`)
- **SubDataset** — smaller version suitable for prototyping, debugging, and benchmarking (branch: `sub`)

This HuggingFace repository uses **branches** to separate large data files and metadata:

- `main` → metadata, scripts, annotations
- `full` → FullDataset
- `sub` → SubDataset

---

## Dataset Structure

```
SynLiDAR/
├── FullDataset/
│   ├── sequences/
│   │   ├── 00.zip
│   │   ├── …
│   │   └── 12.zip
│   └── readme.txt

├── SubDataset/
│   ├── sequences/
│   │   ├── 00.zip
│   │   ├── …
│   │   └── 12.zip
│   └── readme.txt

├── annotations.yaml
├── read_data.py
└── README.md
```

> ⚠️ Some large sequences have been split into chunks (e.g. `06_part01.zip`, `06_part02.zip`, …) to avoid exceeding file size limits.
> These parts are functionally equivalent to the original `06.zip`.

---

## Contents

- `sequences/*.zip` — Each zip contains LiDAR frames from a single drive sequence
- `annotations.yaml` — Semantic categories and label mappings
- `read_data.py` — Example Python loader to read `.bin` point cloud files
- `readme.txt` — Original dataset notes

---

# 🔽 How to Download the Dataset

## Install huggingface_hub

```bash
pip install huggingface_hub
```
## Download Full Dataset (branch: full)

```python
from huggingface_hub import snapshot_download

path = snapshot_download(
    repo_id="AR-X/SynLiDAR",
    repo_type="dataset",
    revision="full",
    local_dir="./SynLiDAR",  # specify your desired path
)
print(path)
```

## Download Sub Dataset (branch: sub)

```python
from huggingface_hub import snapshot_download

path = snapshot_download(
    repo_id="AR-X/SynLiDAR",
    repo_type="dataset",
    revision="sub",
    local_dir="./SynLiDAR",  # specify your desired path
)
print(path)
```
## 🔧 Merging Split ZIP Files

Some sequences in the full branch were too large (>50GB) and are stored as multiple parts:

Example:
```
06_part01.zip
06_part02.zip
...
06_part11.zip
```

These parts correspond to a single original archive. You can merge them into a single folder using Python:

```python
from pathlib import Path

sequence_dir = Path("SynLiDAR/FullDataset/sequences")
# find bases that have part files
bases = sorted({p.stem.split("_part")[0] for p in sequence_dir.glob("*_part*.zip")})

for base in bases:
    parts = sorted(sequence_dir.glob(f"{base}_part*.zip"))
    output_zip = sequence_dir / f"{base}.zip"

    print(f"[{base}] merging {len(parts)} parts -> {output_zip}")
    with open(output_zip, "wb") as out:
        for p in parts:
            with open(p, "rb") as f:
                out.write(f.read())

print("All sequences merged.")
```


## Citation
```
@inproceedings{xiao2022transfer,  
  title={Transfer learning from synthetic to real lidar point cloud for semantic segmentation},  
  author={Xiao, Aoran and Huang, Jiaxing and Guan, Dayan and Zhan, Fangneng and Lu, Shijian},  
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},  
  volume={36},  
  number={3},  
  pages={2795--2803},  
  year={2022}  
}
```


⚠️ **Disclaimer**: This dataset is intended for research and educational usage.
Make sure to respect local regulations when training or deploying autonomous driving systems.