[Back]

M. McCutchan:
"Implementation of a graph based geometry simplification";
Poster: AGIT 2017, Salzburg; 2017-07-05 - 2017-07-07.

A standard problem in visualizing geographic data at different scales is the simplification of geometries. This issue requires to reduce data of a geometry but also to maintain its substantial representation (see figure 1). This poster presents the implementation of a simplification algorithm transforming geometries into a feature space, based on[1], enabling to determine which vertices of a geometry contribute more significantly to the geometries appearance than other vertices. The feature space is established by an acyclic graph. Based on shortest path calculations the algorithm detects the optimal vertices which should be included in the final result. The algorithm enables to model any geometric aspects which are considered important for the simplification procedure, such as directions or distances between vertices. This ability, to model any aspect of the geometry, excels well established simplification algorithms such as the Douglas-Peucker algorithm[2] or the Reumann-Witkam algorithm[3], as they only consider a fixed scope of geometric aspects in their procedures. The code allows the user to define the impact of four chosen geometric aspects in the simplification, namely: (1) The Euclidean distance between any two vertices, (2) directions between a vertex and its neighboring vertices, (3) the number of vertices between any two vertices, and, (4) the overlapping area of the bounding box defined by any two vertices and the original geometry. Additionally, a degree of simplification can be defined, yielding a higher or lower amount of vertices retained. The program provides this functionality using a graphical user interface (GUI) which communicates with the backend. All functionalities are developed in Java. The program generates a simplification of a chosen input geometry, renders as well as displays it and can optionally export it. Thus, the user can see the results of the simplification ad-hoc. This work finalizes with an investigation and illustration of different simplifications of a polygon as well as a line geometry. These simplifications are based on different combinations of the four available geometric aspects as well as different degrees on simplifications.

A standard problem in visualizing geographic data at different scales is the simplification of geometries. This issue requires to reduce data of a geometry but also to maintain its substantial representation (see figure 1). This poster presents the implementation of a simplification algorithm transforming geometries into a feature space, based on[1], enabling to determine which vertices of a geometry contribute more significantly to the geometries appearance than other vertices. The feature space is established by an acyclic graph. Based on shortest path calculations the algorithm detects the optimal vertices which should be included in the final result. The algorithm enables to model any geometric aspects which are considered important for the simplification procedure, such as directions or distances between vertices. This ability, to model any aspect of the geometry, excels well established simplification algorithms such as the Douglas-Peucker algorithm[2] or the Reumann-Witkam algorithm[3], as they only consider a fixed scope of geometric aspects in their procedures. The code allows the user to define the impact of four chosen geometric aspects in the simplification, namely: (1) The Euclidean distance between any two vertices, (2) directions between a vertex and its neighboring vertices, (3) the number of vertices between any two vertices, and, (4) the overlapping area of the bounding box defined by any two vertices and the original geometry. Additionally, a degree of simplification can be defined, yielding a higher or lower amount of vertices retained. The program provides this functionality using a graphical user interface (GUI) which communicates with the backend. All functionalities are developed in Java. The program generates a simplification of a chosen input geometry, renders as well as displays it and can optionally export it. Thus, the user can see the results of the simplification ad-hoc. This work finalizes with an investigation and illustration of different simplifications of a polygon as well as a line geometry. These simplifications are based on different combinations of the four available geometric aspects as well as different degrees on simplifications.

Geometrie, Simplification, Vereinfachung, GIS, Grafik, Visualisierung