Scientists are now turning to satellite images to locate people living in poverty. In a study published in the journal Science, researchers from the Stanford University used machine learning - the science of designing computer algorithms that learn from data - to extract information about poverty from high-resolution satellite imagery, Xinhua news agency reported.
They found the newly developed approach was able to "make fairly accurate predictions" of impoverished areas across five African countries: Nigeria, Tanzania, Uganda, Malawi, and Rwanda.
"Our method, which requires only publicly available data, could transform efforts to track and target poverty in developing countries," the researchers said in their paper.
Overall, surveys are costly, infrequent, and cannot always reach countries or regions within countries, for instance, due to armed conflict.
Recent studies showed that satellite data capturing nightlights can be used to predict wealth in a given area.
However, nightlight data alone is not effective at differentiating between regions at the bottom end of the income distribution, where satellite images appear uniformly dark.
To circumvent this problem, the new study turned its attention to daylight imagery, which offers higher resolution and can capture features such as paved roads and metal roofs, markers that can help distinguish poor and ultra-poor regions.
The researchers then developed a sophisticated learning algorithm that categorizes these features and found that this method did a surprisingly good job predicting the distribution of poverty.
Compared with nightlight models, the new model was 81 per cent more accurate in predicting poverty in regions under the poverty line, and 99% more accurate in areas that are two times below the poverty line.
"We demonstrate an accurate, inexpensive, and scalable method for estimating consumption expenditure and asset wealth from high-resolution satellite imagery," the study said.
"It also demonstrates how powerful machine learning techniques can be applied in a setting with limited training data, suggesting broad potential application across many scientific domains."