search

feevos / resuneta

mxnet source code for the resuneta semantic segmentation models
Other
123 stars 30 forks source link

extracting field boundaries #9

Open long123524 opened 3 years ago

long123524 commented 3 years ago

Hey @feevos. I am very interested in your article "Deep learning on edge: extracting field boundaries from satellite images with a convolutional neural network". I want to learn more, but I didn’t find your code for this article. Can you please What can I learn from? I am a scientific researcher engaged in remote sensing image classification.

feevos commented 3 years ago

Hi @long123524 , thank you for your interest in our work. This repository has the exact same deep learning semantic segmentation model that was used for the paper you refer too. There was no additional repository for the post processing steps (unfortunately - we apologize for this). However, from the outputs of the resunet-a algorithm (segmentation extent, boundaries, distance transform), you can relatively easily create instance segmentation predictions that can be later translated to shapefiles. There exist tutorials (e.g.) on opencv on how to do that.

You can find perhaps additional information here, where a team of researchers from EO reproduced our work independently.

Kind regards

feevos commented 3 years ago

@long123524 see https://github.com/feevos/resuneta/issues/11 for more information - new paper + code on field boundaries.