Skip to content

The project uses Unet-based improved networks to study Remote sensing image semantic segmentation, which is based on keras.

Notifications You must be signed in to change notification settings

TachibanaYoshino/Remote-sensing-image-semantic-segmentation

Repository files navigation

Remote-sensing-image-semantic-segmentation

The project uses Unet-based improved networks to study Remote sensing image semantic segmentation, which is based on keras.
This project has been used in the Sparse Representation and Intelligent Analysis of 2019 Remote Sensing Image competition.


Requirements

  • python 3.6.8
  • tensorflow-gpu 1.8
  • Keras 2.2.4
  • opencv-python
  • tqdm
  • numpy
  • glob
  • argparse
  • matplotlib
  • tifffile
  • pyjson
  • Pillow 6.0
  • scikit-learn

Usage

1. Download dataset

Link,key:1d4x

2. Create new labels

python create_train_val_label.py

3. Train

eg. python train6_6.py --model checkpoint6_6

4. Download pre-trained weights

Link

5. Test

eg. python test.py --model 'checkpoint6_6'+ '/' + 'weights-039-0.7205-0.8099.h5'

Results

Since the original remote sensing image is too large, a partial screenshot of the test results is given here.

About

The project uses Unet-based improved networks to study Remote sensing image semantic segmentation, which is based on keras.

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages