2010 IEEE International Geoscience and Remote Sensing Symposium, 2010
Automatic classification of plant diseases at an early stage is vital for precision crop protecti... more Automatic classification of plant diseases at an early stage is vital for precision crop protection. Our aim was to identify sugar beet leaves inoculated with Cercospora beticola before symptoms are visible. Therefore hyperspectral reflection between 400 and 1050 nm was observed. Relevant wavelengths have to be found in order to implement practical sensor systems with reduced development costs. The main contribution of this study is to identify a minimal subset which is sufficient for separating healthy and inoculated leaves. The heuristic of Hall which analyses the relevance of a feature subset considering the intercorrelation among the features was applied. In order to select a good subset in a reasonable amount of time a genetic algorithm was used. This way enabled a subset of only seven out of 462 wavelengths, which nevertheless enabled us to identify low disease severity ≤ 5% with a classification accuracy of 84.3%. Disease severity above 5% was classified with 99.8%.
Proceedings of the 17th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems - GIS '09, 2009
... Dover Publications. [2] Gold, C. 2003. But is it GIS? Journal of Geospatial Engi-neering, 5(2... more ... Dover Publications. [2] Gold, C. 2003. But is it GIS? Journal of Geospatial Engi-neering, 5(2):11-26. [3] Gröger, G., Kolbe, TH, Czerwinski, A., and Nagel, C. 2008. OpenGIS City Geography Markup Language (CityGML) Encoding Standard. Version 1.0.0, OGC Doc. No. 08-007r1. ...
Page 1. Knowledge-Based Design of Ergonomic Lighting for Underground Scenarios I Wolfram Burgard ... more Page 1. Knowledge-Based Design of Ergonomic Lighting for Underground Scenarios I Wolfram Burgard and Stefan Lfittringhaus-Kappel and Lutz Pltimer Institut f'tir Informatik III, Universi~t Bonn, Rtmerstr. 164, W-5300 Bonn 1 ... [3] Sterling, L. Logical Levels of Problem Solving. ...
Proceedings of the eleventh ACM international symposium on Advances in geographic information systems - GIS 2003, 2003
ABSTRACT This article provides a formal data model to cover many 3D appli-cations in GIS. It is b... more ABSTRACT This article provides a formal data model to cover many 3D appli-cations in GIS. It is based on a 2D model and preserves the algo-rithmic and conceptual simplicity of the 2D case as much as possi-ble. Integrity axioms guarantee the consistency of the model in a ...
... Page 10. ration for the research proposal. 2.1 Goals ... The simultaneity of the eval-uation ... more ... Page 10. ration for the research proposal. 2.1 Goals ... The simultaneity of the eval-uation process forced and at the same tim e allowed to structure the proposals of the joint project, though the individual research groups were responsible for their own proposal. ...
Abstract: Um der Frage nachzugehen, inwieweit eine manuell durchgeführte Bonitur diskreter Objekt... more Abstract: Um der Frage nachzugehen, inwieweit eine manuell durchgeführte Bonitur diskreter Objekte in der Lage ist, die Realität adäquat in einer Verteilungskarte abzubilden, wurden im Rahmen dieser Studie Simulationen mit einer neuen Software durchgeführt. ...
Map labeling is one of the most time-consuming and complex tasks during map generation. Additiona... more Map labeling is one of the most time-consuming and complex tasks during map generation. Additionally, if real-time map labeling is demanded like for screen maps, traditional automatic labeling approaches are not sufficient since they also do not support user interaction methods like zooming and scrolling -change of scale and map clip. Thus, we developed a new approach considering these specific demands for screen maps. The core of our approach is the modeling of possible label conflicts. This includes the identification of such conflicts depending upon their scale of occurrence and the storage of this information. Therefore, we designed a purpose-build data-structure, the so-called reactive conflict graph. This data-structure is built up before a user interacts with the system in the so-called preprocessing phase. It will be exploited during the interaction phase and ensures real-time labeling in a cartographic adequate way. The developed approach supports the labeling of point, line and area objects. The feasibility is successfully demonstrated in a component-based prototype.
Proceedings of the ninth ACM international symposium on Advances in geographic information systems - GIS '01, 2001
Connecting several locations, for example towns by roads, is a typical network planning problem. ... more Connecting several locations, for example towns by roads, is a typical network planning problem. A first approach would be to connect pairs of nodes (network language term for location) with a straight edge. The result is a connected network not considering the total network length and the average trip length. The construction of new junctions, so called Steiner Points, for reducing the total network length and thereby the global costs is more or less known. On the other hand slopes, obstacles or other unsteady costs can not be taken into account during construction. Therefore an independent and homogeneous representation for costs and network planning is necessary.
The sustainable management and continuation of virtual 3D city models puts specific demands on 3D... more The sustainable management and continuation of virtual 3D city models puts specific demands on 3D geodatabases. These demands result from the need to manage three different dimensions of multiple representation: Levels-of-detail, concurrent planning versions, and history over time. In this paper we propose a concept for the integration of all three aspects in the framework of spatial relational database management
The automatic reconstruction of roofs and buildings has been an important research topic during t... more The automatic reconstruction of roofs and buildings has been an important research topic during the last years. This paper elaborates on the automatic and semi-automatic reconstruction of roofs from aerial laser scan data. Two approaches are considered in detail. First, a program has been developed, which segments laser scan data into planar patches, labels them according to their adjacency, and
In this paper we present a concept for the reconstruction of composite objects from LIDAR 3D poin... more In this paper we present a concept for the reconstruction of composite objects from LIDAR 3D point clouds with focus on symmetric and partly recursive structures. The concept is top-down and bases on a generative model given by an attribute grammar. It uses the well known random sample consensus paradigm for model selection. The attribute grammars formalism is used for the modeling of 3D objects. It models composition by production rules and it can also incorporate constraints on parameters and geometric dependencies between parts by attributes. The constraints which are given by the grammar advise to preprocess the data in order to improve the RANSAC algorithm. In the preprocessing step the normal vector of each plane defined by a point and his n neighbors is estimated. The samples are drawn from a subset of these 'needles'.
This paper introduces archetypal dictionaries for a self-taught learning framework for the applic... more This paper introduces archetypal dictionaries for a self-taught learning framework for the application of landcover classification. Self-taught learning, an unsupervised representation learning method, is exploited to learn low-dimensional and discriminative higher-level features, which are used as input into a classification algorithm. Experiments are conducted using a multi-spectral Landsat 5 TM image of a study area in the north of Novo Progresso located in South America. Our results confirm that self-taught learning with archetypal dictionaries provide features, which can be used as input into a linear logistic regression classifier. The obtained classification accuracies are comparable to kernel-based classifier using the original features.
In this paper we show how attribute grammar formalism can be used as a model-ling language for th... more In this paper we show how attribute grammar formalism can be used as a model-ling language for three dimensional geometric models. We focus on 3D building models and their decomposition into building parts. The specification of aggregation and the inheritance of shared parameters is discussed. The focus is on geometrical, topological and semantical con-straints which govern aggregation. We show that whereas aggregation is a natural concept within context-free grammars, inheritance of parameters and constraints can neatly be expressed by semantic rules of attribute grammars. Semantic rules are used to specify form parameters and to inherit location parameters, to encode spatial operations, to represent symmetry and to spec-ify topological constraints. We give an XML schema for spatial attribute grammars, an editing and processing tool and illustrating example results.
2010 IEEE International Geoscience and Remote Sensing Symposium, 2010
Automatic classification of plant diseases at an early stage is vital for precision crop protecti... more Automatic classification of plant diseases at an early stage is vital for precision crop protection. Our aim was to identify sugar beet leaves inoculated with Cercospora beticola before symptoms are visible. Therefore hyperspectral reflection between 400 and 1050 nm was observed. Relevant wavelengths have to be found in order to implement practical sensor systems with reduced development costs. The main contribution of this study is to identify a minimal subset which is sufficient for separating healthy and inoculated leaves. The heuristic of Hall which analyses the relevance of a feature subset considering the intercorrelation among the features was applied. In order to select a good subset in a reasonable amount of time a genetic algorithm was used. This way enabled a subset of only seven out of 462 wavelengths, which nevertheless enabled us to identify low disease severity ≤ 5% with a classification accuracy of 84.3%. Disease severity above 5% was classified with 99.8%.
Proceedings of the 17th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems - GIS '09, 2009
... Dover Publications. [2] Gold, C. 2003. But is it GIS? Journal of Geospatial Engi-neering, 5(2... more ... Dover Publications. [2] Gold, C. 2003. But is it GIS? Journal of Geospatial Engi-neering, 5(2):11-26. [3] Gröger, G., Kolbe, TH, Czerwinski, A., and Nagel, C. 2008. OpenGIS City Geography Markup Language (CityGML) Encoding Standard. Version 1.0.0, OGC Doc. No. 08-007r1. ...
Page 1. Knowledge-Based Design of Ergonomic Lighting for Underground Scenarios I Wolfram Burgard ... more Page 1. Knowledge-Based Design of Ergonomic Lighting for Underground Scenarios I Wolfram Burgard and Stefan Lfittringhaus-Kappel and Lutz Pltimer Institut f'tir Informatik III, Universi~t Bonn, Rtmerstr. 164, W-5300 Bonn 1 ... [3] Sterling, L. Logical Levels of Problem Solving. ...
Proceedings of the eleventh ACM international symposium on Advances in geographic information systems - GIS 2003, 2003
ABSTRACT This article provides a formal data model to cover many 3D appli-cations in GIS. It is b... more ABSTRACT This article provides a formal data model to cover many 3D appli-cations in GIS. It is based on a 2D model and preserves the algo-rithmic and conceptual simplicity of the 2D case as much as possi-ble. Integrity axioms guarantee the consistency of the model in a ...
... Page 10. ration for the research proposal. 2.1 Goals ... The simultaneity of the eval-uation ... more ... Page 10. ration for the research proposal. 2.1 Goals ... The simultaneity of the eval-uation process forced and at the same tim e allowed to structure the proposals of the joint project, though the individual research groups were responsible for their own proposal. ...
Abstract: Um der Frage nachzugehen, inwieweit eine manuell durchgeführte Bonitur diskreter Objekt... more Abstract: Um der Frage nachzugehen, inwieweit eine manuell durchgeführte Bonitur diskreter Objekte in der Lage ist, die Realität adäquat in einer Verteilungskarte abzubilden, wurden im Rahmen dieser Studie Simulationen mit einer neuen Software durchgeführt. ...
Map labeling is one of the most time-consuming and complex tasks during map generation. Additiona... more Map labeling is one of the most time-consuming and complex tasks during map generation. Additionally, if real-time map labeling is demanded like for screen maps, traditional automatic labeling approaches are not sufficient since they also do not support user interaction methods like zooming and scrolling -change of scale and map clip. Thus, we developed a new approach considering these specific demands for screen maps. The core of our approach is the modeling of possible label conflicts. This includes the identification of such conflicts depending upon their scale of occurrence and the storage of this information. Therefore, we designed a purpose-build data-structure, the so-called reactive conflict graph. This data-structure is built up before a user interacts with the system in the so-called preprocessing phase. It will be exploited during the interaction phase and ensures real-time labeling in a cartographic adequate way. The developed approach supports the labeling of point, line and area objects. The feasibility is successfully demonstrated in a component-based prototype.
Proceedings of the ninth ACM international symposium on Advances in geographic information systems - GIS '01, 2001
Connecting several locations, for example towns by roads, is a typical network planning problem. ... more Connecting several locations, for example towns by roads, is a typical network planning problem. A first approach would be to connect pairs of nodes (network language term for location) with a straight edge. The result is a connected network not considering the total network length and the average trip length. The construction of new junctions, so called Steiner Points, for reducing the total network length and thereby the global costs is more or less known. On the other hand slopes, obstacles or other unsteady costs can not be taken into account during construction. Therefore an independent and homogeneous representation for costs and network planning is necessary.
The sustainable management and continuation of virtual 3D city models puts specific demands on 3D... more The sustainable management and continuation of virtual 3D city models puts specific demands on 3D geodatabases. These demands result from the need to manage three different dimensions of multiple representation: Levels-of-detail, concurrent planning versions, and history over time. In this paper we propose a concept for the integration of all three aspects in the framework of spatial relational database management
The automatic reconstruction of roofs and buildings has been an important research topic during t... more The automatic reconstruction of roofs and buildings has been an important research topic during the last years. This paper elaborates on the automatic and semi-automatic reconstruction of roofs from aerial laser scan data. Two approaches are considered in detail. First, a program has been developed, which segments laser scan data into planar patches, labels them according to their adjacency, and
In this paper we present a concept for the reconstruction of composite objects from LIDAR 3D poin... more In this paper we present a concept for the reconstruction of composite objects from LIDAR 3D point clouds with focus on symmetric and partly recursive structures. The concept is top-down and bases on a generative model given by an attribute grammar. It uses the well known random sample consensus paradigm for model selection. The attribute grammars formalism is used for the modeling of 3D objects. It models composition by production rules and it can also incorporate constraints on parameters and geometric dependencies between parts by attributes. The constraints which are given by the grammar advise to preprocess the data in order to improve the RANSAC algorithm. In the preprocessing step the normal vector of each plane defined by a point and his n neighbors is estimated. The samples are drawn from a subset of these 'needles'.
This paper introduces archetypal dictionaries for a self-taught learning framework for the applic... more This paper introduces archetypal dictionaries for a self-taught learning framework for the application of landcover classification. Self-taught learning, an unsupervised representation learning method, is exploited to learn low-dimensional and discriminative higher-level features, which are used as input into a classification algorithm. Experiments are conducted using a multi-spectral Landsat 5 TM image of a study area in the north of Novo Progresso located in South America. Our results confirm that self-taught learning with archetypal dictionaries provide features, which can be used as input into a linear logistic regression classifier. The obtained classification accuracies are comparable to kernel-based classifier using the original features.
In this paper we show how attribute grammar formalism can be used as a model-ling language for th... more In this paper we show how attribute grammar formalism can be used as a model-ling language for three dimensional geometric models. We focus on 3D building models and their decomposition into building parts. The specification of aggregation and the inheritance of shared parameters is discussed. The focus is on geometrical, topological and semantical con-straints which govern aggregation. We show that whereas aggregation is a natural concept within context-free grammars, inheritance of parameters and constraints can neatly be expressed by semantic rules of attribute grammars. Semantic rules are used to specify form parameters and to inherit location parameters, to encode spatial operations, to represent symmetry and to spec-ify topological constraints. We give an XML schema for spatial attribute grammars, an editing and processing tool and illustrating example results.
Uploads
Papers by Lutz Plümer