An Approach for the Classification of Urban Building Structures Based on Discriminant Analysis Techniques

Recognition of urban structures is of interest in cartography and urban modelling. While a broad range of typologies of urban patterns have been published in the last century, relatively little research on the automated recognition of such structures exists. This work presents a sample‐based approach for the recognition of five types of urban structures: (1) inner city areas, (2) industrial and commercial areas, (3) urban areas, (4) suburban areas and (5) rural areas. The classification approach is based only on the characterisation of building geometries with morphological measures derived from perceptual principles of Gestalt psychology. Thereby, size, shape and density of buildings are evaluated. After defining the research questions we develop the classification methodology and evaluate the approach with respect to several aspects. The experiments focus on the impact of different classification algorithms, correlations and contributions of measures, parameterisation of buffer‐based indices, and mode filtering. In addition to that, we investigate the influence of scale and regional factors. The results show that the chosen approach is generally successful. It turns out that scale, algorithm parameterisation, and regional heterogeneity of building structures substantially influence the classification performance.     

Geomorphometric analysis of a rocky coastline: an example from Hornsund,Svalbard

We describe a method to extract a number of quantitative morphometric characteristics of a rocky coastline. The algorithm, which is implemented in the free open-source software GNU Octave, takes as its input an ordered set of coordinate pairs defining the coastline. It finds vertices of specific angularity at two different levels, first approximating convex features (headlands) and then looking for local maxima (embayment edges) in their vicinity. It extracts numerical parameters for each embayment delimited by the former points, including basic geometric parameters (e.g. area, depth and width) and coefficients expressing shapes of embayments (e.g. form factor, coefficient of asymmetry and compactness). These we adapted from morphometric studies of other geomorphic features. Complex headlands and random sectors of the coastline which are delimited by embayment edges are automatically removed. The method was developed and tested on Hornsund in Svalbard, where 20 headlands, 28 embayment edges and 20 embayments were identified, but could be adapted to a wide range of coastal situations. We consider a range of possible applications of the methods, including to the classification of coastlines, change detection and identifying controls on their morphologies, and we discuss the choice of optimum parameters for the algorithm.     

Land use changes in Central Kenya from the 1950s — A possibility to generalise?

Land use changes in Kenya's central Highlands were examined with the purpose of comparing findings from Murang'a District with a new study carried out in a similar environment, Nyeri District. In addition, a generalisation of the findings was examined and methods were analysed to investigate its use in other areas. Aerial photographs were used to analyse land use and soil and water conservation in form of terraces. Furthermore, farmers were interviewed in order to obtain information about how they perceive environmental changes and how these changes have affected their livelihoods. Interpretation of aerial photographs together with field verification and interviews give information about environmental changes and their effects. Population censuses were investigated and findings were compared. Results from this study show that the two study areas have similar physical conditions with decreasing soil fertility. The main difference is the cultivated cash crops, tea and coffee in Nyeri, but only coffee in Murang'a. Also, more land is terraced in Nyeri than in Murang'a. The reason why there are adequate terraces in Nyeri is because of the multitude of cash crops that require terraces for their establishment. Compared with Murang'a, farmers in Nyeri are more satisfied with their situation, mainly due to regular payment from tea and a lower population pressure. This study shows that generalisation of the findings can be made but in order to extrapolate it further to the whole central Highlands more studies and knowledge about the whole area are required. The study also shows that methods used to examine environmental changes can be used elsewhere. Although the recommendations are to generalise with caution so that the final results are reliable and true.     

The CartACom model: transforming cartographic features into communicating agents for cartographic generalisation

Our research is concerned with automated generalisation of topographic vector databases in order to produce maps. This article presents a new, agent-based generalisation model called CartACom (Cartographic generalisation with Communicating Agents), dedicated to the treatment of areas of low density but where rubber sheeting techniques are not sufficient because some eliminations or aggregations are needed. In CartACom, the objects of the initial database are modelled as agents, that is, autonomous entities, that choose and apply generalisation algorithms to themselves in order to increase the satisfaction of their constraints as much as possible. The CartACom model focuses on modelling and treating the relational constraints, defined as constraints that concern a relation between two objects. In order to detect and assess their relational constraints, the CartACom agents are able to perceive their spatial surroundings. Moreover, to make the good generalisation decisions to satisfy their relational constraints, they are able to communicate with their neighbours using predefined dialogue protocols. Finally, a hook to another agent-based generalisation model – AGENT – is provided, so that the CartACom agents can handle not only their relational constraints but also their internal constraints. The CartACom model has been applied to the generalisation of low-density, heterogeneous areas like rural areas, where the space is not hierarchically organised. Examples of results obtained on real data show that it is well adapted for this application.