The problems and progress made in the development of automated methods for the recognition and extraction of urban land use features from the LANDSAT digital data of the UK are described. In the data examined so far density slicing in one waveband appears to be almost as effective as multiwaveband classification in selecting the urban areas. While absolute boundaries may be difficult to delineate, new developments may be recognisable which would be adequate for monitoring needs. It appears important to include the effects of seasonal change and make use of texural analysis in the classification process.

Date of this Version