https://github.com/LiheYoung/SenseEarth2020-ChangeDetection

1st place solution to the Satellite Remote Sensing Image Change Detection Challenge hosted by SenseTime
https://github.com/LiheYoung/SenseEarth2020-ChangeDetection

change-detection remote-sensing-image satellite-images semantic-segmentation

Last synced: 8 days ago
JSON representation

1st place solution to the Satellite Remote Sensing Image Change Detection Challenge hosted by SenseTime

• Host: GitHub
• URL: https://github.com/LiheYoung/SenseEarth2020-ChangeDetection
• Owner: LiheYoung
• License: mit
• Created: 2020-10-19T05:57:55.000Z (about 4 years ago)
• Default Branch: master
• Last Pushed: 2021-12-08T03:52:22.000Z (almost 3 years ago)
• Last Synced: 2024-08-02T02:11:07.814Z (3 months ago)
• Topics: change-detection, remote-sensing-image, satellite-images, semantic-segmentation
• Language: Python
• Homepage: https://rs.sensetime.com
• Size: 6.94 MB
• Stars: 281
• Watchers: 4
• Forks: 53
• Open Issues: 4
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-remote-sensing-change-detection - LiheYoung/SenseEarth2020-ChangeDetection

README

        
# SenseEarth2020 - ChangeDetection

**1st place in the Satellite Remote Sensing Image Change Detection 

[Challenge](https://rs.sensetime.com/competition/index.html#/info) 

hosted by [SenseTime](https://www.sensetime.com/cn).**

## Our Method

### Task Description

Given two images of the same scene acquired at different times, we are required to mark the changed 

and unchanged areas. Moreover, as for the changed areas, we need to annotate their detailed semantic masks. 

The change detection task in this competition can be decomposed into two sub-tasks:

* binary segmentation of changed and unchanged areas.

* semantic segmentation of changed areas.

### Model

![image](https://github.com/LiheYoung/SenseEarth2020-ChangeDetection/blob/master/docs/pipeline.png)

### Pseudo Labeling

**The core practice is using self-distillation strategy to assign pseudo labels to unchanged areas.**

Specifically, in our experiments, predictions of five HRNet-based segmentation models are ensembled, 

serving as pseudo labels of unchanged areas. 

The overall training process can be summarized as:

* Training multiple large segmentation models.

* Ensembling their predictions on unchanged areas.

* Training a smaller model with both labeled and pseudo labeled areas.

For more details, please refer to the 

[technical report](https://github.com/LiheYoung/SenseEarth2020-ChangeDetection/blob/master/docs/technical%20report.pdf) 

and [presentation](https://github.com/LiheYoung/SenseEarth2020-ChangeDetection/blob/master/docs/presentation.pptx).

## Getting Started

### Dataset

[Description](https://rs.sensetime.com/competition/index.html#/data) | [Download [password: f3qq]](https://pan.baidu.com/s/1Yg90vlAiKezSoxH7WEoV6g) 

### Pretrained Model

[HRNet-W18](https://drive.google.com/file/d/1LHVq88dh4uRiq7pGbb-ZFoP8Na-XRUU2/view?usp=sharing) | [HRNet-W40](https://drive.google.com/file/d/1kyjB_FwYqTwe4B1Cj8my0SLQ9eUQA8nn/view?usp=sharing) | [HRNet-W44](https://drive.google.com/file/d/1ZHFHSrsL-Qoq-fzS7XoPayyQ_20V3RZ0/view?usp=sharing) | [HRNet-W48](https://drive.google.com/file/d/1o9ATkRkD2rVm4MMXY7HP3zNVWdB_jd0I/view?usp=sharing) | [HRNet-W64](https://drive.google.com/file/d/18aZW7Kja2WnzDT0IsvFAGUQktgv81Zd4/view?usp=sharing)

### Final Trained Model

[PSPNet-HRNet-W18](https://drive.google.com/file/d/1xRoltAyFjT7DxvuxS5WGdjG0YOlIG8i6/view?usp=sharing) | [PSPNet-HRNet-W40](https://drive.google.com/file/d/1QKQ_gAoyA_ysFWj5kDO_TlPnGJKOUyYf/view?usp=sharing)

### File Organization

```

# store the whole dataset and pretrained backbones

mkdir -p data/dataset ; mkdir -p data/pretrained_models ;

# store the trained models

mkdir -p outdir/models ; 

# store the pseudo masks

mkdir -p outdir/masks/train/im1 ; mkdir -p outdir/masks/train/im2 ;

# store predictions of validation set and testing set

mkdir -p outdir/masks/val/im1 ; mkdir -p outdir/masks/val/im2 ;

mkdir -p outdir/masks/test/im1 ; mkdir -p outdir/masks/test/im2 ;

├── data

    ├── dataset                    # download from the link above

    │   ├── train                  # training set

    |   |   ├── im1

    |   |   └── ...

    │   └── val                    # the final testing set (without labels)

    |

    └── pretrained_models

        ├── hrnet_w18.pth

        ├── hrnet_w40.pth

        └── ...

```

### Training

```

# Please refer to utils/options.py for more arguments

# If hardware supports, more backbones can be trained, such as hrnet_w44, hrnet_w48

CUDA_VISIBLE_DEVICES=0,1,2,3 python train.py --backbone hrnet_w18 --pretrained --model pspnet --lightweight

```

### Pseudo Labeling & Re-training 

```

# This step is optional but important in performance improvement

# Modify the backbones, models and checkpoint paths in L20-40 in label.py manually according to your saved models

# It is better to ensemble multiple trained models for pseudo labeling

# Pseudo labeling

CUDA_VISIBLE_DEVICES=0,1,2,3 python label.py

# Re-training

CUDA_VISIBLE_DEVICES=0,1,2,3 python train.py --backbone hrnet_w18 --pretrained --model pspnet --lightweight --use-pseudo-label

```

### Testing

```

# Modify the backbones, models and checkpoint paths in L39-44 in test.py manually according to your saved models

# Or simply use our final trained models

CUDA_VISIBLE_DEVICES=0,1,2,3 python test.py

```