search

SpaceNetChallenge / utilities

Packages intended to assist in the preprocessing of SpaceNet satellite imagery data corpus to a format that is consumable by machine learning algorithms.
Other
248 stars 97 forks source link

createDataSpaceNet.py doesn't convert all objects in geojson to bounding boxes for Darknet #81

Closed reachsumit closed 5 years ago

reachsumit commented 5 years ago

I am trying to generate Darknet style bounding boxes for spacenet tif files. The .gejson file for the corresponding tif files has more than 15 buildings. However, the createDataSpaceNet.py always outputs one entry in the text file. No error is given in the terminal.

ashnair1 commented 5 years ago

@reachsumit I'm facing a similar problem while working with the Spacenet Off Nadir dataset. Converting annotations to Darknet format results in only a few annotations being converted and those that are converted are empty files. Were you able to successfully convert your annotations to Darknet annotations?

reachsumit commented 5 years ago

@ash1995 IIRC, it didn't work for me. So, I used this: https://medium.com/the-downlinq/getting-started-with-spacenet-data-827fd2ec9f53 with a bit of customization.