Note that this hasn't been updated in a while, hopefully I will return to update it soon
Run the following commands to clone this repo and SuperGlue's repo:
TODO: git setup
$ git clone https://github.com/OSUPCVLab/VehicleGeopositioning.git
$ cd VehicleGeopositioning
$ git clone https://github.com/magicleap/SuperGluePretrainedNetwork.git
Run the following commands to set up the dependancies:
$ pip3 install -r requirements.txt
- Get a google maps API key. Use the following link to set up the api key. Make sure you copy it (do not share API keys, each person should use their own) https://developers.google.com/maps/documentation/embed/get-api-key TODO: more directions here when navigating the google API
- Open the
keys.py
document and paste in your API key into theGOOGLE_API_KEY
field
main.py
there is the script which runs everything. To run in simpily typepython3 main.py
into the command line.- Here is how to use custom data. There are two predefined arguments
--framesDir
and--dataDir
. Currently there are defined to point to the sample data. The convention is that if an imagedata/images/img.png
exists, then there is a correspondingdata/parameters/img.txt
that supplies meta data for the image. Each image's meta data should provide the same information in the same format as currently exists in the sample data. ex.GPS = 40.01265351729508,-83.00969233386985 height = 70 rotation = 0