GeoProcessor system is a portable computer environment for intelligent support of research and problem solving on space-time geodata analysis. GeoProcessor environment provides user with tools for a wide range of requirements: from data screening and general data analysis to sophisticated complex investigation and decision making on geological exploration and seismic hazard assessment problems.


Java 1.1 ! Please use Netscape 4.06 or Internet Explorer 4.0 and upper

North-East China


Northern Caucasia

Word data center online
Project number: RBRF 99-07-90326

Network Geoinformation Systems for Presentation and Analysis of Spatio-Temporal Information in Earth and Human Dimension Sciences
Project number: RBRF 00-07-90100

in cooperation with UIPE RAS, GC RAS, CPI Rosaviakosmos, China Seismological Bureau

Merapi volcano


Turkey region

SPIN!

Spatial Mining for Data of Public Interest

Information Societies Technology Programme

Contract number: EU IST-10536 SPIN!

Michael May (co-ordinator) E-mail:Michael.May@gmd.de

Kresna region


W. Corinthos Gulf and Thiva-Oropos region

ASPELEA

ASSESMENT OF SEISMIC POTENTIAL IN EUROPEAN LARGE EARTHQUAKE AREAS

EC Project INCO-COPERNICUS

Contract number: IC-15CT-97-0200

Gerassimos Papadopoulos (co-ordinator) E-mail:g.papad@egelados.gein.noa.gr

Central Europe region

Data from CD-ROM "The European Geotraverse Database"


Head of the project: Dr.Sc. Valeri Gitis gitis@iitp.ru
Java programming: Boris Osher, Anton Chekhov, Ivan Denissovitch


  1. Overwiew
  2. Cartographic representation of raster, vector and point data
    1. Composing a map
    2. Changing map settings
    3. Reading values of raster fields
    4. Visualization of cross-section by user defined profile
    5. Compiling similarity with precedent for raster fields
  3. Data transformation
    1. Calculation of raster fields with algebraic and logical operations
    2. Formation samples of precedents
    3. Building topography model
    4. Raster field filtering
    5. Event filtering
    6. Line filtering
  4. Plausible reasoning facilities
    1. Computing similarity on a set of precedents
    2. Computing membership for two classes
    3. Computing certainty for classification on monotone feature space
    4. Computing nonparametric regression on a set of precedents

Geoprocessor is information technology for on-line cartographic presentation, intelligent modeling and complex analysis of space-time properties of geological environment via Internet.

The system is intended for various users with wide range of requirements: from acquainting with the complex of geodata on a region of interest to sophisticated geoinformation problems.

Geoprocessor can be used for decision support in environmental monitoring, natural hazard assessment, natural and man made disaster mitigation, natural resources exploration.

1. Cartographic representation of raster, vector and point data

Composing a map

In GeoProcessor a map is a set of point, line and area objects and raster fields, combined in raster and vector layers. E.g., information about cities, rivers, countries or earthquake epicenters is visualized on the map as different instances of vector layers, continued data like topography, gravity or magnetic anomalies are mapped as raster layers. Along with the database of prepared to use thematic maps, the system provides mechanism to compose new maps from arbitrary set of raster and vector layers.

To display one of prepared thematic map, select its name from the Map List and press the button Show.

To compose new map you need to take following steps:

  1. Press the button New map in the main window.
  2. Specify name of the map in the field Name of appeared dialog.
  3. Select the raster layers to be displayed from the Fields list.
  4. Select the vector layers to be displayed from the Vectors list.
  5. Press the button Add Layer (>) to include selected layers into the map.
  6. Change order of the map layers (if it needed) with the button Order (^) *.
  7. Specify visible layers with the button Show/Hide Layer (+/-) **.
  8. To add co-ordinate grid press the button Add Grid (#).
  9. Press the button OK to add new map to the Map List.
  10. Display the map as described above.

* In GeoProcessor layers have to be ordered in particular manner: raster layers -> vector layers and co-ordinate grid.

** Alternative way to change order and visibility of map layers is to choose View > Layers. The Layers Panel appears on the left side of the map window. It contains checkboxes to show/hide vector layers and drop down list to change display order of raster fields.

To modify the content of any existing map display it and choose Map > Compose. Add or remove map components, change its order and visibility with the Edit Map Layers dialog as described above.

Changing map settings

GeoProcessor allow users to change map scale, coordinates of displayed region and parameters of coordinate grid.

To change scale of the map, choose View > Zoom In, Zoom Out or Zoom to Fit. Map scale will be changed without interpolation of raster layers. To interpolate the map after zooming or resizing map window, choose View > Refresh.

To change coordinates of displayed region choose View > Region and specify values of latitude and longitude in appeared dialog.

To change parameters of coordinate grid choose View > Grid and specify coordinates of covered region and step of grid cells.

Reading values of raster fields

To read a value in arbitrary point of raster field choose Tools > Values and click inside a map to specify the point of interest. If the map consists of several raster layers, values of all rasters in specified point will be calculated and displayed in the Values in a Point window.

Visualization of cross-section by user defined profile

The system provides calculating and visualization of cross-section for several raster layers selected for the map. Choose Tools > Crossection to start entering the profile. Press mouse button inside a map and move pointer to specify segment of profile line. Release mouse button to fix the segment. Double click left mouse button to complete entering the profile. Cross-section window shows graphics of all selected raster fields along the profile. Field names and its values are displayed below the graphic by corresponding colors. Move mouse cursor on the graphic to examine different profile points. Cross cursor on the map indicates position of corresponding points.

Compiling similarity with precedent for raster fields

Cartography precedent analysis realized in the GeoProcessor supports compiling the map of similarity with specified point by a set of features. Similarity is calculated in the metrics l2 and c with the values of raster fields normalized by its variances. To compile similarity choose Tools > Analysis. Select fields to be processed and metric in the Similarity Analysis dialog. Specify the value of R that gives a maximal distance from the sample points in the normalized feature space to select the definition area of similarity function. Press the button OK. Double click left mouse button inside the map to specify the point of the precedent. This point is marked with blue triangle. The map of similarity with the precedent will be compiled and displayed in a grayscale. More similar areas are darker while less similar are lighter. Double click left mouse button again to add next precedent point and rebuild the map. To clear the map of similarity click the right mouse button anywhere inside the map. To cancel analysis choose Tools > Analysis again to uncheck the option.

 

2. Data transformation

Calculation of raster fields with algebraic and logical operations

GeoProcessor provides facilities to explore initial data with the set of algebraic and logical functions that can be applied to raster fields. The system realizes the following functions:

To calculate new field on the basis of initial set of raster data choose Transform > Field combination. Specify the name of the field in the dialog Input Function. Select the name of the function to be applied from the Functions list. Select field(s) to be processed from the Fields list. You can combine several functions with arithmetic and logical operators from the Operators list. Modify entered formulae manually (if it needed) and press the button OK. New field will be calculated, added to the map and displayed.

Formation samples of precedents

Sample is an aggregate of map objects used in operations of plausible inference on a complex of raster maps. Samples can be loaded from file or entered manually. GeoProcessor lets you create two types of samples: sample of points and sample of areas. Point sample contains arbitrary set of points marked on map, while sample of areas includes all points inside specified polygon(s).

To create sample of points choose New Sample item from Map menu. The Edit Panel appears at the top of the window. Press one of the buttons with shape pictogram on Edit Panel to select the type of marker. Click anywhere inside a map to enter a point of sample. To remove a point from the sample press Select button and click the point to be deleted. Remove selected point(s) by the button Delete. To specify a value in a point of the sample select the point and enter the value in the input field of Edit Panel, then press Enter.

To create sample of areas choose Map > New Areas. Press the button Enter Polygon to begin. Press mouse button inside a map and move pointer to specify segment of polygon. Release mouse button to fix the segment. Close the polygon with the button Close. All points inside the polygon will be processed as sample. To remove the polygon, select it and press the button Delete.


Building topography model field

Topography model is 3D-like representation of topography or other surface.

To build topography model field on the basis of initial raster field choose Transform > Topo model. Specify a light source position (azimut, angle with horizon) and initial field multiply factor. Press <Preview> button to view preliminary results of transformation, <OK> button to preserve resulting field and <Cancel> button to exit without saving transformation results.

 

Raster field filtering

The next options are available for raster field filtering:

Mean - smooth a field with a square window 2R*2R. The result field value is equal:

( integral of field values in the square)/(area size)

Median - median smoothing of initial field values within the window.

R.M.S. - field of standard deviations within a square 2R*2R.

Local anomalies - calculating the differences between the values of initial field and the mean value.

Gradient module - building field of absolute values of the gradient vector.

Gradient azimuth - calculating the field of azimuth values of the gradient vector.

Maximum - calculate maximum values within a square 2R*2R.

Minimum - calculate minimum values within a square 2R*2R.

Maximum - Minimum -calculate a difference(maximum-minimum) values within a square 2R*2R.

To apply raster field filter choose Transform > Field. Specify resulting field Name, filtering type (mean, median, RMS, local anomalies, gradient module, gradient azimuth, maximum, minimum, maximum-minimum), radius and initial field. Press <Preview> button to view preliminary results of transformation, <OK> button to preserve resulting field and <Cancel> button to exit without saving transformation results.

 

Event filtering

The next options are available for creating raster field from events:

Density - field of density of events within the circle of user-defined radius R. The value rerult field is equal to (the number of events within the window)/3.14*R*R.

Power - field of weighted density of events in the window. The value in a field grid knot is equal to (S of values of events within the window)/3.14*R*R.

Distance - field of minimum distances from the raster points to the events. The value of a field grid knot is equal to the distance from the knot to the nearest event within the circle of radius R. If there is no events within the circle, then the value is equal R.

Closeness - field of closeness to the events. Resulting field value is equal to (1-r/R), where r is the distance defined as in the previous option.

Influence - field of density of events calculated with triangular window. The result field value is equal to S (1-r(n)/R) , were r(n) is the distance from the knot to the all events within the circle of user-defined radius R.

To build field by event filtering choose Transform > Event Filter. Specify resulting field Name, filtering type (density, power, distance, closeness or influence), radius and event vector layer. Press <Preview> button to view preliminary results of transformation, <OK> button to preserve resulting field and <Cancel> button to exit without saving transformation results.

 

Line filtering

The next options are available for creating raster field from lines:

Distance - distances from raster point to the nearest line within user-defined radius R.

Influence - the nearness is calculated by formula S (1 - r/R), where r - distance and R - radius, as above. The sum is calculating for all lines crossing the circle of user-defined radius R.

Nearness - the nearness is calculated by formula (1 - r/R), where r - distance and R - radius, as above.

Number - (number of the lines within the circle with radius R)/3.14*R^2.

Sum Length - total length of the lines within the circle with radius R.

To build field by line filtering choose Transform > Line Filter. Specify resulting field Name, filtering type (distance, influence, nearness, number or sum length), radius and line vector layer. Press <Preview> button to view preliminary results of transformation, <OK> button to preserve resulting field and <Cancel> button to exit without saving transformation results.

 

3.Plausible reasoning facilities

Computing similarity on a set of precedents

The similarity is a function of the distance from point under consideration to the nearest sample point in feature space. Similarity function is equal one at the sample point and decreases up to zero at the distance R (in normalized feature space) from the sample points.

To calculate similarity, first you need to enter the sample of precedents. It may be sample of points or areas entered manually or loaded from the file. After the sample entering do the following:

  1. Choose Transform > Similarity Field.
  2. Specify the name of similarity field in the Similarity Field Creation dialog.
  3. Select the metric and specify the value of R.
  4. Select the name of the sample from the list of samples or areas.
  5. Press the button Add (>).
  6. Select the field(s) to be processed from the Fields list.
  7. Press the button OK.

The field of similarity with the precedent will be computed and displayed in a grayscale. Areas that more similar to specified precedent set are darker while less similar are lighter.

Computing membership for two classes

To compute membership field you need to prepare two samples: sample A and sample B. Membership function is equal 1 at the sample points of class A and equal -1 at the sample points of class B.After the samples creating do the following:

  1. Choose Transform > Membership field.
  2. Specify the name of the membership field in the Membership Field Creation dialog.
  3. Select the metric and specify the value of R.
  4. Select the sample A from the Samples (Areas) list.
  5. Press the upper button Add (>).
  6. Select the sample B from the Samples (Areas) list.
  7. Press the lover button Add (>).
  8. Press the button OK.

The field of membership will be computed and displayed in a grayscale. Areas nearer to the sample A are darker while areas nearer to the sample B are lighter.

Computing certainty for classification of monotone feature space

To compute the field of monotony, enter the sample of precedents. After the sample entering do the following:

  1. Choose Transform > Monotony Field.
  2. Specify the name of monotony field in appeared dialog.
  3. Select the name of the sample from the list of samples or areas.
  4. Press the button Add (>).
  5. Select the field(s) to be processed from the Fields list.
  6. Press the button OK.

The field of monotony will be computed and displayed in a grayscale. The darkness of result field shows a number of precedents for which monotony condition on the feature space is confirmed.

Computing regression on a set of precedents

To compute regression field enter the sample of points and specify the value of each point. After the sample entering do the following:

  1. Choose Transform > Regression Field.
  2. Specify the name of regression field in the dialog.
  3. Select the metric and specify the value of R.
  4. Select the name of the sample from the list of samples or areas.
  5. Press the button Add (>).
  6. Select the field(s) to be processed from the Fields list.
  7. Press the button OK.

The regression field will be computed and displayed. The resulting field values are weighted-mean precedent point values. The weights are inversely proportional to similarity value.