Predicting Wealth Score from Remote Sensing Satellite Images and Household Survey Data Using Deep Learning

Shashank Shekhar; Pratibha Singh; Rashmi Mishra; Sunil Kumar

doi:10.24423/cames.2024.722

Shashank Shekhar Edunet Foundation, Gurugram, India
Pratibha Singh Department of Computer Science & Engineering, Krishna Engineering College, Ghaziabad, India
Rashmi Mishra Delhi Technological University, India
Sunil Kumar Institute for Communication Systems, University of Surrey, United Kingdom

DOI: http://dx.doi.org/10.24423/cames.2024.722

Abstract

The most exigent call of the United Nations’ 17 sustainable goals is to end poverty everywhere by 2030. Unlike in the past, when poverty was measured based on data collected through ground-level surveys, the new technology adopted by many developing and developed countries is to estimate the poverty index using remote sensing satellite images with the help of machine learning techniques. Our approach demonstrates the prediction of cluster wealth score and establishes the relationship between wealth score obtained from Demographic and Health Survey (DHS) data and remote sensing satellite images of India by calculating Pearson’s correlation coefficient (r²). The implementation results have been analyzed in four phases. Phase 1 comprises four regression models (RMs): Ridge, RANSAC, Lasso, and k-nearest neighbor for feature extraction from a pre-trained convolutional neural network model using daylight & nightlight images. Here, the Lasso RM outperforms the others and is best suited for predicting the wealth score. Phase 2 categorizes daylight images with DHS data, where the Lasso RM efficiently generates the cluster wealth score. Phase 3 focuses on images of specific regions of Delhi, Tamil Nadu, Maharashtra and Telangana, using the Lasso RM, as it emerged as the best predictor of cluster wealth score in the previous two phases. Phase 4 compares the results attained through our proposed model with existing results.

Keywords

convolutional neural network, Demographic and Health Survey Data, Inception V3, transfer learning,

References

1. World Data Lab, World Poverty Clock, https://worlddata.io/portfolio/world-povertyclock/.
2. D. Eugenie, R. Kennedy, J. Urpelainen, Satellite data for the social sciences: measuring rural electrification with night-time lights, International Journal of Remote Sensing, 39(9): 2690–2701, 2018, doi: 10.1080/01431161.2017.1420936.
3. J. Neal, M. Burke, M. Xie, W.M. Davis, D.B. Lobell, S. Ermon, Combining satellite imagery and machine learning to predict poverty, Science, 353(6301): 790–794, 2016, doi: 10.1126/science.aaf789.
4. NOAA, Nightlight Images, DMSP-OLS Dataset, https://ngdc.noaa.gov/eog/dmsp/.
5. The DHS Program, Available datasets, https://dhsprogram.com/data/available-datasets.cfm.
6. K.N. Nischal, R. Radhakrishnan, S. Mehta, S. Chandani, Correlating night-time satellite images with poverty and other census data of India and estimating future trends, [in:] CODS’15: Proceedings of the 2nd ACM IKDD Conference on Data Sciences, New York, NY, United States, pp. 75–79, Association for Computing Machinery (ACM), 2015, doi: 10.1145/2732587.2732597.
7. United Nations, A World That Counts: Mobilizing the Data Revolution for Sustainable Development – Report of the Secretary-General’s Independent Expert Advisory Group on the Data Revolution for Sustainable Development, 2014, https://digitallibrary.un.org/record/3882725.
8. United Nations, Sustainable Development Goals: 17 Goals to Transform Our World, 2015, http://www.un.org/sustainabledevelopment/sustainable-development-goals/.
9. A. Ajami, M. Kuffer, C. Persello, K. Pfeffer, Identifying a slums’ degree of deprivation from VHR images using convolutional neural networks, Remote Sensing, 11(11): 1282, 2019, doi: 10.3390/rs11111282.
10. Daylight Images, Google HERE Maps API, www.here.com.
11. Living Standard Measurement Survey, The World Bank, www.worldbank.org/lsms.
12. E. Sheehan, C. Meng, M. Tan, B. Uzkent, N. Jean, M. Burke, D. Lobell, S. Ermon, Predicting economic development using geolocated Wikipedia articles, [in:] Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, pp. 2698–2706, Association for Computing Machinery, New York, NY, United States, 2019, doi: 10.1145/3292500.3330784.
13. S. Pandey, T. Agarwal, N.C. Krishnan, Multi-task deep learning for predicting poverty from satellite images, [in:] Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 32(1), pp. 7793–7798, New Orleans, Lousiana, USA, 2018, doi: 10.1609/aaai.v32i1.11416.
14. L. Duan, T. Hu, E. Cheng, J. Zhu, C. Gao, Deep convolutional neural networks for spatiotemporal crime prediction, [in:] Proceedings of the International Conference on Information and Knowledge Engineering (IKE), H.R. Arabnia [Ed.], pp. 61–67, The Steering Committee of The World Congress in Computer Science, Computer Engineering and Applied Computing (WorldComp), 2017.
15. L. Guie, Z. Cai, X. Liu, J. Liu, S. Shiliang, A comparison of machine learning approaches for identifying high-poverty counties: Robust features of DMSP/OLS night-time light imagery, International Journal of Remote Sensing, 40(15): 5716–5736, 2019, doi: 10.1080/01431161.2019.1580820.
16. R. Ruchir, Temporal poverty prediction in developing countries, Stanford CS229: Machine Learning, Stanford University, California, 2017, https://cs229.stanford.edu/proj2017/final-reports/5244050.pdf.
17. S.P. Subash, R.R. Kumar, K.S. Aditya, Satellite data and machine learning tools for predicting poverty in rural India, Agricultural Economics Research Review, 31(2): 231–240, 2018, doi: 10.5958/0974-0279.2018.00040.X.
18. B. Raj, A simple guide to the versions of the inception network, Towards Data Science, May 29, 2018, https://towardsdatascience.com/a-simple-guide-to-the-versions-ofthe-inception-network-7fc52b863202 [accessed: Aug. 20, 2020].
19. M. Xie, N. Jean, M. Burke, D. Lobell, S. Ermon, Transfer learning from deep features for remote sensing and poverty mapping, [in:] Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 30(1), pp. 3929–3935, AAAI Press, Palo Alto, California USA, 2016, doi: 10.1609/aaai.v30i1.9906.
20. G.R. Watmough, C.L.J. Marcinko, C. Sullivan, K. Tschirhart, P.K. Mutuo, C.A. Palm, J.C. Svenning, Socioecologically informed use of remote sensing data to predict rural household poverty, [in:] Proceedings of the National Academy of Sciences, Vol. 116(4), pp. 1213–1218, 2019, doi: 10.1073/pnas.1812969116.
21. A. Perez, S. Ganguli, S. Ermon, G. Azzari, M. Burke, D. Lobell, Semi-supervised multitask learning on multispectral satellite images using Wasserstein generative adversarial networks (GANs) for predicting poverty, arXiv, 2019, arXiv: 1902.11110.
22. S. Piagessi et al., Predicting City Poverty Using Satellite Imagery, [in:] Proceedings of the IEEE/CFV Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pp. 90–96, Long Beach, California, USA, 2019.
23. J.E. Steele et al., Mapping poverty using mobile phone and satellite data, Journal of the Royal Society Interface, 14(127): 20160690, 2017, doi: 10.1098/rsif.2016.0690.
24. G. Catamuro, A. Muhebwa, J. Taneja, Street smarts: measuring intercity road quality using deep learning on satellite imagery, [in:] Proceedings of the 2nd ACM SIGCAS Conference on Computing and Sustainable Societies, Association for Computing Machinery, pp. 145–154, New York, NY, United States, 2019, doi: 10.1145/3314344.3332493.
25. P.K. Suraj, A. Gupta, M. Sharma, S.B. Paul, S. Banerjee, On monitoring development indicators using high resolution satellite images, arXiv, 2017, arXiv: 1712.02282.
26. ImageNet, Database, http://www.image-net.org.
27. S. Bhattacharyya, Ridge and Lasso regression: L1 and L2 regularization, Towards Data Science, 2018, https://towardsdatascience.com/ridge-and-lasso-regression-a-complete-guidewith-python-scikit-learn-e20e34bcbf0b.
28. RANSAC Regression, Regression Models, 2018, https://scikit-learn.org/stable/modules/classes.html.
29. KNN Nearest Neighbors Regression, 2018, https://scikit-learn.org/stable/supervised_learning.html#supervised-learning.
30. Pearson’s Correlation Coefficient, Correlation Coefficient, 2018, https://en.wikipedia.org/wiki/Pearson_correlation_coefficient.
31. E. Sheehan, Z. Nabulsi, C. Meng, Utilizing Latent Embeddings of Wikipedia Articles to Predict Poverty, Preprint, Stanford University, 2018, https://cs229.stanford.edu/proj2018/report/134.pdf [accessed: Aug. 20, 2020].