Pyro-Scrapper is a specialized tool designed for scraping images from Alert Wildfire, splitting these images by camera, and preparing them for future analysis using Pyro-Engine (not yet developed). The primary objective of this project is to augment the Pyronear dataset with images of wildfires, which can be either actual fires or false positives, both of which are valuable for analysis.
-
Image Scraping: Downloads images from Alert Wildfire, leveraging the
dl_images.py
script. This involves fetching camera data, processing images, and handling various states and timezones. View Script -
Image Splitting: The
split_cams.py
script is used to split images based on the camera viewpoint -
Data Processing by Pyro-engine: We pass all images into pyro-engine as if they came from a Pyronear camera in order to make predictions and send alerts via our api.
-
Scraping Images: The script
dl_images.py
downloads images from Alert Wildfire, categorizing them based on the camera's state and source. It processes the images to ensure they meet the required standards (e.g., removing grayscale images). -
Splitting by Camera:
split_cams.py
TO BE DEVELOPED -
Data Processing by Pyro-engine: TO BE DEVELOPED
Certainly! Here's the revised Usage section with the added code snippets:
- Clone the repository.
- Ensure you have the required dependencies by installing them from
requirements.txt
.
Install requirements
pip install -r requirements.txt
To download images:
python src/dl_images.py
To split images by camera:
python/split_cams.py
These commands will initiate the image scraping and view point splitting.
Please refer to CONTRIBUTING
if you wish to contribute to this project.
This project is developed and maintained by the repo owner and volunteers from Pyronear.
Distributed under the Apache 2 License. See LICENSE
for more information.