/ChangeOS

ChangeOS: Building damage assessment via Deep Object-based Semantic Change Detection - (RSE 2021)

Primary LanguagePython

Building damage assessment for rapid disaster response with a deep object-based semantic change detection framework:
from natural disasters to man-made disasters

by Zhuo Zheng, Yanfei Zhong, Junjue Wang, Ailong Ma and Liangpei Zhang

[Paper] [BibTeX]



This is an official implementation of ChangeOS in our RSE 2021 paper Building damage assessment for rapid disaster response with a deep object-based semantic change detection framework: from natural disasters to man-made disasters.

Highlights

  • Deep object-based semantic change detection framework (ChangeOS) is proposed.
  • ChangeOS seamlessly integrates object-based image analysis and deep learning.
  • City-scale building damage assessment can be achieved within one minute.
  • A global-scale dataset is used to evaluate the effectiveness of ChangeOS.
  • Two local-scale datasets are used to show its great generalization ability.

Getting Started

Installation

pip install changeos

Requirements:

  • pytorch == 1.10.0
  • python >=3.6
  • skimage
  • Pillow

Usage

# changeos has four APIs
# (e.g., 'list_available_models', 'from_name', 'visualize', 'demo_data')
import changeos


# constructing ChangeOS model
# support 'changeos_r18', 'changeos_r34', 'changeos_r50', 'changeos_r101'
model = changeos.from_name('changeos_r101') # take 'changeos_r101' as example

# load your data or our prepared demo data
# numpy array of shape [1024, 1024, 3], [1024, 1024, 3]
pre_disaster_image, post_disaster_image = changeos.demo_data()

# model inference
loc, dam = model(pre_disaster_image, post_disaster_image)

# put color map on raw prediction
loc, dam = changeos.visualize(loc, dam)

# visualize by matplotlib
import matplotlib.pyplot as plt
plt.subplot(121)
plt.imshow(loc)
plt.subplot(122)
plt.imshow(dam)
plt.show()

Citation

If you use ChangeOS in your research, please cite the following paper:

@article{zheng2021building,
  title={Building damage assessment for rapid disaster response with a deep object-based semantic change detection framework: from natural disasters to man-made disasters},
  author={Zheng, Zhuo and Zhong, Yanfei and Wang, Junjue and Ma, Ailong and Zhang, Liangpei},
  journal={Remote Sensing of Environment},
  volume={265},
  pages={112636},
  year={2021},
  publisher={Elsevier}
}