README.md

Maker Dataset

Code for creating dataset from drone video. Result: drone image == image from satellite map

File log_reader - parse log.file of drone for receive: number of frame with coordinates. Return dict

File video_reader - cuts video into frames (you can set the step between frames). For each frame from the drone, it generates a request to Google api based on information from the log file (which we get from log_reader). Receives in response an image from Google maps. Saves the image

It is also worth noting that for each frame from the drone, the function generates 5 frames from Google maps with an up-down-right-left shift and the original.

Dependencies

python 3.10

GOOGLE_API_KEY

Usage

if __name__ == '__main__':
    path_to_log = "PATH/TO/LOG.SRT"
    log_info = log_proccesing(path_to_log)
    path_to_video = "PATH/TO/VIDEO.MP4"
    path_dron = 'PATH/TO/DRONE_IMAGES/SAVE'
    path_map = 'PATH/TO/MAPS_IMAGES/SAVE'
    apiKey = "GOOGLE_API_KEY"
    video_proccesing(log_info, path_to_video, path_dron, path_map,apiKey)

Run

Clone the project

  git clone https://github.com/tidehackathon/team-next-cop.git

Go to the project directory

  cd team-next-cop/MakerDataset

Create enviroment

  python3 -m venv env

Install dependencies

  pip install -r req.txt

Command for start

  python video_reader.py