Points Classification

Owned by Tore Sannes (Unlicensed)
Last updated: Oct 31, 2014
7 min read

In this section:

Introduction

This command will generate classificated points for the selected chart(s).
Input to the classification algorithms are the grids called Seafloor that are located under each "chart name" folder.

 


The Points Classification dialog


The content of the menu

The classification options are: Shoals, Deeps, Saddles, Ridges and Valleys

The points ar defined as:

  • Shoal point: The shallowest point (node or sounding) in an area. All surrounding nodes are deeper.
  • Deep point: The deepest point (node or sounding) in an area. All surrounding nodes are shallower.
  • Saddle point: The shallowest point (node) in a fairway. The surrounding nodes are deeper on two sides and shallower on the two perpendicular sides.
  • Ridge point: A break point (node) on a ridge. The surrounding nodes are deeper on three sides and shallower on the fourth side.
  • Valley point: A break point (node) in a valley. The surrounding nodes are shallower on three sides and deeper on the fourth side.

 

Classification points

The result is a set of points that tells where the selected classification points are located. The classification points are saved under the <chart name> folder and a classification folder with the name of the option and the parameters selected.

The parameters are set in the four tabs: Shoals/Deeps, Saddles, Ridges and Valleys

When the correct parameters for the selected options  are set, click Execute to perform the points classification.

 

The Shoals/Deeps tab



Specify settings for shoal objects

Specify how to find those grids nodes that are the shallowest within the Minimum distance. The parameters are:

  • Mask width (grid cells): Is a mask moving over the grid and checking the centre node with the nodes around. Default value is 2.
  • Minimum distance (m): A shoal is detected when the centre node is shallower than the rest of the nodes in the mask. After the detection the points are sorted on increasing depth values, and by starting with the first (and shallowest) point, all points within a distance smaller than the Minimum distance will be removed. Then the same test is performed for the next valid point, etc. Default value is 25.

The shoals are saved as <chart name> / classification / shoals_<mask width>_<distance>.

 

Specify settings for deep objects

Specify how to find those grids nodes that are the deepest within the Minimum distance. The parameters are:

  • Mask width (grid cells): Is a mask moving over the grid and checking the centre node with the nodes around. Default value is 2.
  • Minimum distance (m): A deep is detected when the centre node is deeper than the rest of the nodes in the mask. After the detection the points are sorted on decreasing depth values, and by starting with the first (and deepest) point, all points within a distance smaller than the Minimum distance will be removed. Then the same test is performed for the next valid point, etc. Default value is 25.

The deeps are saved as <chart name> / classification / deeps_<mask width>_<distance>.

 

Find corresponding Shoals/Deeps in Soundings

This option enables a search in the Soundings dataset for the nearest point that was found in the Seafloor grid. The corresponding points are saved as <chart name> / classification / <type>_in_soundings_<mask width>_<distance>.

 

The Saddles tab

Spscify how to find those grids nodes that have two low and two high neighbour points. The parameters are:

  • Mask width (grid cells): A filter moving over the grid and checking the centre node with the nodes around. Default value is 2. 
  • Minimum depth difference (m): The minimum depth difference between the actual node and the neighboring topography. Default value is 0.1.

The saddle points are found by searching the grid nodes along the x and y axis, and then the diagonals within the mask. The criteria for a saddle point is that along one of the directions, the neighboring terrain should be shallower in both directions, while in the perpendicular axis the terrain should be deeper in both directions.

The search directions for saddles, ridges and valleys for mask width 2 (5x5).

After the detection the points are sorted on increasing depth values, and by starting with the first (and shallowest) point, all points within a distance smaller than the Mask width will be removed. All the remaining points ar checked the same way.

The result dataset is saved as <chart name> / classification / saddles_<mask width>.

 

The Ridges tab

Will find those grids nodes that define a ridge. The parameters are:

  • Mask width (grid cells): A filter moving over the grid and checking the centre node with the nodes around. Default value is 2.
  • Slope angle towards shallower (deg): The vertical angle towards shallower terrain. This angle should not be too steep for a ridge point. Default value is 2.
  • Slope angle towards deeper (deg): The vertical angle towards shallower terrain. This angle should be steeper than the upward angle. Default value is 5.
  • Maximum contour angle (deg): The angle of a contour line passing through the selected point. A smaller angle gives a sharper feature. Default value is 120.

The ridge points are found by searching the grid nodes along the x and y axis, and then the diagonals within the mask. The criteria for a ridge point is that along one of the directions the neighboring terrain should be shallower in one direction and deeper in the other, while in the perpendicular axis the terrain should be deeper in both directions.

The slope angles towards shallower and deeper terrain are the test criteria for defining the the limit for the slope.

When a candidate is detected, the next check is the contour angle between the grid node and two interpolated points with the same depth value. If the sharpness of the angle is smaller than the value in the parameter, the node is accepted as a ridge point. 

The contour angle for ridge and valley points.

After the detection the points are sorted on increasing depth values, and by starting with the first (and shallowest) point, all points within a distance smaller than the Mask width will be removed. All the remaining points ar checked the same way.

The result is saved as <chart name> / classification / ridges_'mask width'_'shallower angle'_'deeper angle'_'contour angle'. 

To see all ridge points set Slope angle towards shallower = 1, Slope angle towards deeper = 1 and Maximum contour angle = 180. 

Test and tune the parameters to get the proper ridge points.

 

The Valleys tab

Will find those grids nodes that define a valley (which is the opposite of a ridge). The parameters are: 

  • Mask width (grid cells): A filter moving over the grid and checking the centre node with the nodes around. Default value is 2.
  • Slope angle towards shallower (deg): The vertical angle towards shallower terrain. This angle should be steeper than the downward angle for a valley point. Default value is 5.
  • Slope angle towards deeper (deg): The vertical angle towards shallower terrain. This angle should be not be too steep for a valley point. Default value is 2.
  • Maximum contour angle (deg): The angle of a contour line passing through the selected point. A smaller angle gives a sharper feature. Default value is 120.

The valley points are found by searching the grid nodes along the x and y axis, and then the diagonals within the mask. The criteria for a valley point is that along one of the directions the neighboring terrain should be deeper in one direction and shallower in the other, while in the perpendicular axis the terrain should be shallower in both directions.

The slope angles towards shallower and deeper terrain are the test criteria for defining the the limit for the slope.

When a candidate is detected, the next check is the contour angle between the grid node and two interpolated points with the same depth value. If the sharpness of the angle is smaller than the value in the parameter, the node is accepted as a valley point. 

After the detection the points are sorted on increasing depth values, and by starting with the first (and shallowest) point, all points within a distance smaller than the Mask width will be removed. All the remaining points ar checked the same way.

The result is saved as <chart name> / classification / valleys_'maskwidth'_'shallower angle'_'deeper angle'_'contour angle'.

To see all valley points set Slope angle towards shallower = 1, Slope angle towards deeper = 1 and Maximum contour angle = 180.

Test and tune the parameters to get the proper valley points.

The Results

After executing the point classification a dialog will pop up showing the result from the classifications. The report itself is saved in the project structure in the folder named Reports located in the same folder as the terrain surface. The report itself is named as YYYYMMDD Points Classification (the name of the executed command).

All reports from all runs will be saved in the Reports folder.