2018Journal article

Urban Land-Use Mapping Using a Deep Convolutional Neural Network with High Spatial Resolution Imagery

Huang, B., Zhao, B., & Song, Y.

Remote Sensing of Environment, 214, 73–86

Read publication ↗

Research question

Can deep convolutional models distinguish detailed urban land-use categories from imagery?

Geographic scope

Urban land-use mapping

Evidence base

High-resolution multispectral imagery and deep learning

Approach

A deep convolutional neural network learns spatial and spectral patterns in high-resolution multispectral remote-sensing imagery.

Why it matters

Detailed land-use maps provide an important physical layer for studying urban form, activity and environmental exposure.

Research connections

Where this paper sits in the lab’s research system