Skip to content
Code for satellite image segmentation using Unet or Mask RCNN and comparing these two approaches.
Branch: master
Clone or download
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data Add files via upload Jun 19, 2018
maskRCNN update README Jun 19, 2018
unet fix typo Jun 26, 2018
utils update unet_utils.py Jun 19, 2018
.gitignore uncomment prertrained weights May 25, 2018
LICENSE Initial commit May 9, 2018
LOL_prepare_data.ipynb update README Jun 19, 2018
README.md adding link to blog post Jul 7, 2018

README.md

segmentation-unet-maskrcnn

Code for satellite image segmentation using Unet or Mask RCNN and comparing these two approaches.

Please, see more details in the blog post Satellite Images Segmentation and Sustainable Farming

Get started

LOL_prepare_data.ipynb: data augmentation using flip (mirror effect), rotate90, flip up to down, flip left to right.

Data

input data folder structure for maskRCNN

  • train/valid/test
    • jpg (aerial images)
    • jpg4 (aerial images + hill shade data)
    • polygon (masks)

input data folder structure for unet

  • train/valid/test
    • jpg (aerial images)
    • hill (hill shade data)
    • polygon (masks)
You can’t perform that action at this time.