The use of context in the classification of urban aerial imagery

Urban regions present some of the most challenging areas for the remote sensing community. Many different types of land cover have similar spectral responses, making them difficult to distinguish from one another. Traditional per-pixel classification techniques suffer particularly badly because they only use these spectral properties to determine a class, and no other properties of the image, such as context. This project presents the results of the classification of a deeply urban area of Dudley, West Midlands, using 4 methods: Supervised Maximum Likelihood, SMAP, ECHO and Unsupervised Maximum Likelihood. An accuracy assessment method is then developed to allow a fair representation of each procedure and a direct comparison between them. Subsequently, a classification procedure is developed that makes use of the context in the image, though a per-polygon classification. The imagery is broken up into a series of polygons extracted from the Marr-Hildreth zero-crossing edge detector. These polygons are then refined using a region-growing algorithm, and then classified according to the mean class of the fine polygons. The imagery produced by this technique is shown to be of better quality and of a higher accuracy than that of other conventional methods. Further refinements are suggested and examined to improve the aesthetic appearance of the imagery. Finally a comparison with the results produced from a previous study of the James Bridge catchment, in Darleston, West Midlands, is made, showing that the Polygon classified ATM imagery performs significantly better than the Maximum Likelihood classified videography used in the initial study, despite the presence of geometric correction errors.

The application of remote sensing to the management of urban wildlife habitats

The project set out with two main aims. The first aim was to determine whether large scale multispectral aerial photography could be used to successfully survey and monitor urban wildlife habitats. The second objective was to investigate whether this data source could be used to predict population numbers of selected species expected to be found in a particular habitat type. Panchromatic, colour and colour infra-red, 1:2500 scale aerial photographs, taken in 1981 and 1984, were used. For the orderly extraction of information from the imagery, an urban wildlife habitat classification was devised. This was based on classifications already in use in urban environments by the Nature Conservancy Council. Pilot tests identified that the colour infra-red imagery provided the most accurate results about urban wildlife habitats in the study area of the Blackbrook Valley, Dudley. Both the 1981 and 1984 colour infra-red photographs were analysed and information was obtained about the type, extent and distribution of habitats. In order to investigate whether large scale aerial photographs could be used to predict likely animal population numbers in urban environments, it was decided to limit the investigation to the possible prediction of bird population numbers in Saltwells Local Nature Reserve. A good deal of research has already been completed into the development of models to predict breeding bird population numbers in woodland habitats. These models were analysed to determine whether they could be used successfully with data extracted from the aerial photographs. The projects concluded that 1:2500 scale colour infra-red photographs can provide very useful and very detailed information about the wildlife habitats in an urban area. Such imagery can also provide habitat area data to be used with population predictive models of woodland breeding birds. Using the aerial photographs, further investigations into the relationship between area of habitat and the breeding of individual bird species were inconclusive and need further research.