README.md

# Lane Lines Dataset

This dataset is designed for lane line detection in autonomous driving scenarios. It contains polyline annotations representing visible lane boundaries (center lines, edge lines, adjacent lane markings) captured from three vehicle-mounted cameras across driving sessions in urban and suburban environments.

## Dataset Description

- **Source:** Fleet of autonomous vehicles, urban and suburban roads
- **Sensor:** 3x LUCID TRI054S-CC cameras (2880×1860), front/side facing
- **Coverage:** ~6.4K polyline annotations, ~1.7K unique images, 7 unique lane lines per frame
- **Splits:** Train 70% (4,490 rows, 1,213 images), Test 30% (1,886 rows, 521 images) - split by image to prevent leakage
- **Format:** Parquet (annotations) + JPEG (images) in `train/` and `test/` subdirectories
- **Coordinates:** Pixel-space polyline, each `data` entry contains alternating (x, y) pairs

### Classes (1)

| Class | Count | Description |
|-------|-------|-------------|
| line | 6,376 | Lane boundary polyline (5 points = 10 coordinates) |

### Line IDs

Each annotated line carries an `object_id` (0–6) indicating a specific lane boundary within the scene:

| object_id | Count | Description |
|-----------|-------|-------------|
| 0 | 1,693 | Left ego-lane boundary |
| 1 | 1,686 | Right ego-lane boundary |
| 2 | 1,542 | Adjacent lane boundary |
| 3 | 886 | Far lane boundary |
| 4 | 409 | Additional lane marking |
| 5 | 90 | Distant left boundary |
| 6 | 60 | Distant right boundary |

### Data Fields

| Field | Type | Description |
|-------|------|-------------|
| msg_id | VARCHAR | UUID linking to the original line message |
| data | DOUBLE[10] | Polyline as alternating (x, y) pixel coordinates — 5 points |
| label | VARCHAR | Annotation class (`line`) |
| object_id | BIGINT | Lane line identifier (0–6) |
| timestamp_ns | BIGINT | ROS bag timestamp (nanoseconds) |
| image_path | VARCHAR | Relative path to JPEG in `train/` or `test/` |

### Data Splits

| Split | Rows | Images |
|-------|------|--------|
| train | 4,490 | 1,213 |
| test  | 1,886 | 521   |

### Dataset Structure

```
lines_dataset/
├── annotations.parquet   # All annotations (6,376 rows)
├── train/
│   ├── camera_1C0FAF5250E2/   # 99 images
│   ├── camera_1C0FAF5CA7B6/   # 60 images
│   └── camera_1C0FAF5CC14D/   # 1,054 images
└── test/
    ├── camera_1C0FAF5250E2/   # 51 images
    ├── camera_1C0FAF5CA7B6/   # 30 images
    └── camera_1C0FAF5CC14D/   # 440 images
```

### Usage

```python
import pandas as pd

df = pd.read_parquet("annotations.parquet")
print(f"{len(df)} annotations, {df.image_path.nunique()} unique images")
```