PAIRS is an Image Recognition System powered by Artificial Intelligence, designed for the detection and analysis of pulsar candidates. This project aims to add another filter to the identification of pulsars in astronomical data by leveraging machine learning models.

To get started with PULSAR, follow these steps:

1. API Key Setup:

   • Deploy this Roboflow model to obtain an API key. Then create a config.ini file to store the API key in this format, quotation marks are not used:

     [Credentials]
     ROBOFLOW_API_KEY = API_KEY

2. Installation:

   • Clone this repository to your local machine:

     git clone https://github.com/metalmerge/PULSAR.git
     cd PULSAR

   • Run this command to download all the requirements:

     pip install -r requirements.txt

3. Run the Code:

   • Execute predict_unchecked_JSON.py to run the pulsar candiate model predictor. A JSON file will be returned with all the selected pulsar candidates.

    python3 predict_unchecked_JSON.py

   • Run low_DM_Remover.py to remove any candidates with DM < 1 from found_Pulsar_Candidates.json.

    python3 low_DM_Remover.py

• The JSON_to_images folder is used for converting checked JSON to images to use for training data.
• JSON_input is for the unchecked or checked input JSON data.
• JSON_output is for the output of JSON files.
• known_JSON_Pulsar_Data is for the two folders of known pulsar candidates and RFI.

• Project Lead: Dimitry Ermakov
• Data Source: NANOstars (Wenke Xia)

Special thanks to NANOstars for their willingness in providing the pulsar data and to Wenke Xia for the additional help with procuring the data.

If you have any questions or suggestions, please feel free to contact me:

• Email
• GitHub