Showing papers by "Lars Bernard published in 2012"

Classification-driven air pollution mapping as for environment and health analysis

Abstract: This article presents an air pollution modelling approach and its use in health applications within the EO2HEAVEN project. The model is based on a multidimensional Inverse Distance Weighting and makes use of artificial distances between area attributes. The number and type of attributes used is fully customizable and can be adapted according to specific application fields and data preconditions. It is kept flexible and simple and thus, suitable to be used within a Spatial Data Infrastructure to provide access to realtime air pollution information via the internet. In a prototypical implementation the model is applied to estimate the concentration of particular matter and ozone in the Federal State of Saxony, Germany.

Moving Code - Sharing Geospatial Computation Logic on the Web

Abstract: Software reuse is a common paradigm for building software systems and has been proven to facilitate maintenance and increase productivity. Valuable algorithms are produced in many scientific projects, open source initiatives or commercial software manufacturers. Despite the success, reuse of computational logic in large distributed systems such as Spatial Data Infrastructures (SDI) remains an issue. This paper presents an approach for describing, publishing and sharing geospatial computational logic on the Web as Moving-CodePackages. These packages are service-oriented software components that contain algorithmic code and its description, i.e. the contracted functionality, platform and hardware plus basic information about the exploitation rights. We also provide a generic messaging mechanism to publish processing logic on the Web and to make users aware of new versions and updated algorithms. The suggested approach is expected to provide benefits in data-driven science, cloud computing environments and offers new possibilities to deal with data protection requirements.