|
| |
|
| |
|
|
Bayesian classification of digital images by web application
Talich, M. ; Böhm, O. ; Soukup, Lubomír
The contribution introduces web application for image classification that has been developed at the Research Institute of Geodesy, Topography and Cartography in the framework of grant project InGeoCalc (supported by Ministry of education of the Czech Republic). The web application is aimed to display, examine and classify digital image data. The data are expected to be obtained from Internet by means of Web Map Services (WMS) or from other sources (possibly non-registered). Image data from different sources can be combined and presented as composition of layers (coverage) with adjustable degrees of transparency. After gathering the data, Bayesian (supervised) classification is applied to distinguish separate regions in the image. User can choose between several classification methods and adjust pertinent parameters. Furthermore, several subsequent basic analytical tools are offered, namely computation of distances, areas or perimeters related to the classified regions, simple statistical summaries about classification results (e.g. distribution of classes, percentage of non-classified regions, etc.). The classification results and registration parameters can be saved for further use. The web application is based on common Internet standards (HTML, Javascript, SVG). The only requirement for running the application is an up-to-date Internet browser supporting SVG (Scalable Vector Graphics). Typical usage of the web application can involve land cover mapping based on satellite or aerial images. The application is available free of charge for any Internet user.
|
|
|
Possibilities and Limits of Affine Transformation Application to Georeferencing
Soukup, Lubomír
Affine transform is one of the most common transforms having been applied for several purposes in such branches as geodesy, cartography, GIS, remote sensing, computer graphics etc. Its application in some specific case is, however, limited by requirements on positional precision of transformed points in a given area of interest. This contribution therefore deals with accuracy analysis of affine transform. The aim of the analysis is to provide objective arguments for a decision whether affine transform is suitable for some specific purpose (eg. georeferencing of historical maps, matching of satellite or aerial images). Some generalizations of affine transform are also mentioned, namely method of collocation, Jung's transform, Coons plate, Thin Plate Spline, polynomial transform. Purely probabilistic and statistic approach is applied throughout the accuracy analysis, transformation parameters are estimated by Bayesian approach.
|
|
|
How to cope with positional inaccuracies in old maps?
Soukup, Lubomír
This paper focuses on the utilization of historical essays and their interpretation The thing is that when maps created in various times are compared, position discrepancies are discovered. In such cases it is sometimes difficult to decide whether the discovered discrepancy results from the inaccuracy of old maps or whether a change in the landscape really occured. These problems can be solved by means of the probability theory and mathematical statistics. The mathematical statistics can thus contribute to higher objectiveness of the conclusions of historical and cartographic research.
|
|
| |
|
| |
|
| |
|
| |
|
| |
guest ::
