Deep Population

Deep Learning Approach for Population Estimation
from Satellite Imagery

 Microsoft AI for Earth Awardee


Knowing where people live is a fundamental component of many decision making processes such as urban development, infectious disease containment, evacuation planning, risk management, conservation planning, and more.

To jointly answer the questions of “where do people live” and “how many people live there,” we propose a deep learning model for creating high-resolution population estimations from satellite imagery. Specifically, we train convolutional neural networks to predict population in the USA at a 0.01° x 0.01° resolution grid from 1-year composite Landsat imagery.

We find that aggregating our model’s estimates gives comparable results to the Census county-level population projections and that the predictions made by our model can be directly interpreted, which give it advantages over traditional population disaggregationmethods.

In general, our model is an example of how machine learning techniques can be an effective tool for extracting information from inherently unstructured, remotely sensed data to provide effective solutions to social problems.

One of the results from "A Deep Learning Approach for Population Estimation from Satellite Imagery". This figure shows the images from the testing, unseen, data that are most confidently classified by our model. Notice the types of images that appear from left (roads, few people) to right (buildings, many people) indicate that our deep learning model is learning semantically-relevant features from satellite imagery.


A Deep Learning Approach for Population Estimation from Satellite Imagery
Caleb Robinson, Fred Hohman, Bistra Dilkina
1st ACM SIGSPATIAL Workshop on Geospatial Humanities. Nov 7, 2017. Redondo Beach, USA.

PDF | Slides | Bibtex

Our Team

Caleb Robinson

PhD Student

Fred Hohman

PhD Student

Bistra Dilkina

Assistant Professor

Polo Chau

Assistant Professor