README.md

---
license: cc0-1.0
pretty_name: LEANDRONE_V1
tags:
- line follow
- drone
- vision
size_categories:
- n<1K
viewer: false
---

Repository: https://huggingface.co/datasets/Leandro4002/LEANDRONE_V1  
Download zip: https://public.saraivam.ch/static/LEANDRONE_V1.zip  
Project using this dataset: https://gitlab.com/Leandro4002/drone-follow-line

## Description  

The **LEANDRONE_V1** dataset is a collection of 500 labelled images triying to mimic the photo taken by the front camera of a Bitcraze AI deck 1.1 mounted on a Crazyflie 2.1 nanodrone.
The camera model is a Himax HM01B0 monochrome with dimension 320×320.

This dataset is aimed to create a machine learning model for an autonomous line-following drone.
Each image is labelled with 2 points named "from" and "to" that draw a line in the direction forward.

Each image has random variations of camera angles, line path, light placement and furnitures.

## Example of sample
<img src="render/0.png"/>
<img src="render_label/0.png"/>

```json
[
    {"from": [181, 267], "to": [166, 188]}
]
```

## Crazyflie nanodrone 2.1 and Bitcraze AI deck 1.1:  
<img src="crazyflie_nanodrone_2.1.png" width="300"/>
<img src="bitcraze_aideck1.1.png" width="300"/>

## File structure  
```
|
|
|---render
|   "Black and white images, named from '0.png' to '499.png'"
|
|
|---render_label
|   "Colored images, named from '0.png' to '499.png'.
|   It also contains a red and green sphere, allowing us to viualize
|   the 2 labels 'from'(red sphere) and 'to'(green sphere)"
|
|
|---labels.json
    "json array containing json object with the location of
    the points 'from' and 'to'. The position is an array with
    2 values, the first is x and the second is y. The unit is a pixel.
    x goes from left to right and y goes from top to bottom.
    The position of the object in the array correspond to the name of the image"
```

## Author

Made in Blender 4.0.2 by Leandro SARAIVA MAIA  
20 March 2024  
This work is marked with CC0 1.0 Universal