From Pixels to Progress: Generating Road Network from Satellite Imagery for Socioeconomic Insights in Impoverished Areas

Yanxin Xi,Yu Liu,Zhicheng Liu,Sasu Tarkoma,Pan Hui,Yong Li

2024-06-17

Abstract:The Sustainable Development Goals (SDGs) aim to resolve societal challenges, such as eradicating poverty and improving the lives of vulnerable populations in impoverished areas. Those areas rely on road infrastructure construction to promote accessibility and economic development. Although publicly available data like OpenStreetMap is available to monitor road status, data completeness in impoverished areas is limited. Meanwhile, the development of deep learning techniques and satellite imagery shows excellent potential for earth monitoring. To tackle the challenge of road network assessment in impoverished areas, we develop a systematic road extraction framework combining an encoder-decoder architecture and morphological operations on satellite imagery, offering an integrated workflow for interdisciplinary researchers. Extensive experiments of road network extraction on real-world data in impoverished regions achieve a 42.7% enhancement in the F1-score over the baseline methods and reconstruct about 80% of the actual roads. We also propose a comprehensive road network dataset covering approximately 794,178 km2 area and 17.048 million people in 382 impoverished counties in China. The generated dataset is further utilized to conduct socioeconomic analysis in impoverished counties, showing that road network construction positively impacts regional economic development. The technical appendix, code, and generated dataset can be found at <a class="link-external link-https" href="https://github.com/tsinghua-fib-lab/Road_network_extraction_impoverished_counties" rel="external noopener nofollow">this https URL</a>.

Computer Vision and Pattern Recognition,Artificial Intelligence

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is: **the challenges of accurately monitoring and generating road networks in poverty - stricken areas**, especially how to use satellite images to improve the data integrity and timeliness of road infrastructure in these areas to promote socio - economic development.

### Specific problems include:

1. **Limitations of existing road data**:
- **Incomplete OpenStreetMap (OSM) data**: In poverty - stricken areas, there are a large number of missing OSM data, especially in road connections in remote mountainous areas.
- **Rough and lagging road data provided by official agencies**: These data usually only provide road length information, lacking timeliness and detail.
- **High cost of data from commercial map platforms**: Most researchers and organizations cannot afford to purchase high - quality road data.

2. **Importance of real - time monitoring of road networks**:
- **Reduce the population neglected due to poverty**: Timely and accurate monitoring of road network development is crucial for reducing the population neglected due to poverty.
- **Promote economic and social development**: The development of road infrastructure can improve internal transportation convenience and attract external investment, thereby promoting regional economic and social development.

3. **Advantages of using deep learning and satellite images**:
- **Advantages of high - resolution satellite images**: Satellite images have a long - time span, comprehensive spatial coverage and high spatial resolution, and are an ideal data source for observing roads in poverty - stricken areas.
- **Advancements in deep - learning technology**: In recent years, road - extraction methods based on satellite images have made significant progress, providing a technical basis for solving the above problems.

### Solutions proposed in the paper:

- **Systematic road - extraction framework**: Combine the encoder - decoder architecture and morphological operations to extract road networks from satellite images.
- **Large - scale road - network data set**: Generate a fine - grained road - network data set covering 382 poverty - stricken counties in China, with an area of approximately 794,178 square kilometers and a population of 17.048 million.
- **Economic and social analysis**: Conduct economic and social analysis through the generated road - network data set, verifying the positive impact of road construction on regional economic development (such as GDP).

### Main contributions:

1. Propose a scalable framework for extracting road networks in poverty - stricken counties from satellite images.
2. Generate the first fine - grained road - network data set covering 382 poverty - stricken counties in China.
3. Experimental results show that this framework improves the F1 score by 42.7% compared to the baseline method and can reconstruct approximately 80% of the actual roads.
4. Economic and social analysis indicates that road expansion can quantitatively increase the GDP of poverty - stricken counties by 7%.

Through these efforts, this paper provides important support for road - infrastructure research in the Sustainable Development Goals (SDGs).