Hyperspectral Remote Sensing in Urban Environments

5Classification of Urban Land Use and Land Cover

DOI: 10.1201/9781003329312-5

This chapter will discuss the following topics:

• Classification of urban land use and land cover
  • Image classification
  • Its significance to hyperspectral data
  • Supervised and unsupervised machine learning techniques overview
  • Challenges in land use land cover classification
  • Class codes/models USGS etc.
  • Mathematical preliminaries
• Index-based methods
  • Method intuition, method
  • Different urban indices
• Spectral matching methods
  • Structure – general algo
  • Similarity measures
  • Some of the case studies
• Conventional machine learning (ML, k-NN)
  • Simple image features
  • Proxy features
• Deep learning (with spectral convolutional focus)
  • Spectral, spatial, spectral-spatial features ...