Automated School Location Mapping at Scale from Satellite Imagery Based on Deep Learning
Computer vision for large scale building detection can be very challenging in many environments and settings even with recent advances in deep learning technologies. Even more challenging is modeling to detect the presence of specific buildings (in this case schools) in satellite imagery at a global...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-02-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/14/4/897 |
| _version_ | 1797476810197303296 |
|---|---|
| author | Iyke Maduako Zhuangfang Yi Naroa Zurutuza Shilpa Arora Christopher Fabian Do-Hyung Kim |
| author_facet | Iyke Maduako Zhuangfang Yi Naroa Zurutuza Shilpa Arora Christopher Fabian Do-Hyung Kim |
| author_sort | Iyke Maduako |
| collection | DOAJ |
| description | Computer vision for large scale building detection can be very challenging in many environments and settings even with recent advances in deep learning technologies. Even more challenging is modeling to detect the presence of specific buildings (in this case schools) in satellite imagery at a global scale. However, despite the variation in school building structures from rural to urban areas and from country to country, many school buildings have identifiable overhead signatures that make them possible to be detected from high-resolution imagery with modern deep learning techniques. Our hypothesis is that a Deep Convolutional Neural Network (CNN) could be trained for successful mapping of school locations at a regional or global scale from high-resolution satellite imagery. One of the key objectives of this work is to explore the possibility of having a scalable model that can be used to map schools across the globe. In this work, we developed AI-assisted rapid school location mapping models in eight countries in Asia, Africa, and South America. The results show that regional models outperform country-specific models and the global model. This indicates that the regional model took the advantage of having been exposed to diverse school location structure and features and generalized better, however, the global model was the worst performer due to the difficulty of generalizing the significant variability of school location features across different countries from different regions. |
| first_indexed | 2024-03-09T21:09:01Z |
| format | Article |
| id | doaj.art-3801c63afd9f49638103b2ef1576f77e |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-09T21:09:01Z |
| publishDate | 2022-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-3801c63afd9f49638103b2ef1576f77e2023-11-23T21:53:50ZengMDPI AGRemote Sensing2072-42922022-02-0114489710.3390/rs14040897Automated School Location Mapping at Scale from Satellite Imagery Based on Deep LearningIyke Maduako0Zhuangfang Yi1Naroa Zurutuza2Shilpa Arora3Christopher Fabian4Do-Hyung Kim5Office of Global Innovation, UNICEF, New York, NY 10017, USADevelopment Seed 2, 1226 9th Street NW Second Floor, Washington, DC 20001, USAOffice of Global Innovation, UNICEF, New York, NY 10017, USAOffice of Global Innovation, UNICEF, New York, NY 10017, USAOffice of Global Innovation, UNICEF, New York, NY 10017, USAOffice of Global Innovation, UNICEF, New York, NY 10017, USAComputer vision for large scale building detection can be very challenging in many environments and settings even with recent advances in deep learning technologies. Even more challenging is modeling to detect the presence of specific buildings (in this case schools) in satellite imagery at a global scale. However, despite the variation in school building structures from rural to urban areas and from country to country, many school buildings have identifiable overhead signatures that make them possible to be detected from high-resolution imagery with modern deep learning techniques. Our hypothesis is that a Deep Convolutional Neural Network (CNN) could be trained for successful mapping of school locations at a regional or global scale from high-resolution satellite imagery. One of the key objectives of this work is to explore the possibility of having a scalable model that can be used to map schools across the globe. In this work, we developed AI-assisted rapid school location mapping models in eight countries in Asia, Africa, and South America. The results show that regional models outperform country-specific models and the global model. This indicates that the regional model took the advantage of having been exposed to diverse school location structure and features and generalized better, however, the global model was the worst performer due to the difficulty of generalizing the significant variability of school location features across different countries from different regions.https://www.mdpi.com/2072-4292/14/4/897computer visiondeep learningschool mappinghigh resolution satellite imagery |
| spellingShingle | Iyke Maduako Zhuangfang Yi Naroa Zurutuza Shilpa Arora Christopher Fabian Do-Hyung Kim Automated School Location Mapping at Scale from Satellite Imagery Based on Deep Learning Remote Sensing computer vision deep learning school mapping high resolution satellite imagery |
| title | Automated School Location Mapping at Scale from Satellite Imagery Based on Deep Learning |
| title_full | Automated School Location Mapping at Scale from Satellite Imagery Based on Deep Learning |
| title_fullStr | Automated School Location Mapping at Scale from Satellite Imagery Based on Deep Learning |
| title_full_unstemmed | Automated School Location Mapping at Scale from Satellite Imagery Based on Deep Learning |
| title_short | Automated School Location Mapping at Scale from Satellite Imagery Based on Deep Learning |
| title_sort | automated school location mapping at scale from satellite imagery based on deep learning |
| topic | computer vision deep learning school mapping high resolution satellite imagery |
| url | https://www.mdpi.com/2072-4292/14/4/897 |
| work_keys_str_mv | AT iykemaduako automatedschoollocationmappingatscalefromsatelliteimagerybasedondeeplearning AT zhuangfangyi automatedschoollocationmappingatscalefromsatelliteimagerybasedondeeplearning AT naroazurutuza automatedschoollocationmappingatscalefromsatelliteimagerybasedondeeplearning AT shilpaarora automatedschoollocationmappingatscalefromsatelliteimagerybasedondeeplearning AT christopherfabian automatedschoollocationmappingatscalefromsatelliteimagerybasedondeeplearning AT dohyungkim automatedschoollocationmappingatscalefromsatelliteimagerybasedondeeplearning |