|
Home : EvalViewer : Reference : Tabs :
![]() | Previous: Diagnostics tab |
Lets you work with meshes.
The Make button converts visible cloud points on selected cloud point objects into a fine mesh of small quadrilateral grid elements. Each mesh is a single surface description despite any type of cloud point ordering or cloud overlap.
When you click the Make button, EvalViewer makes a view mesh using input information that you must modify separately beforehand.
To set parameters, use Mesh > Params. The default parameter settings should work well for most situations.
The parameters relevant to the Make View Mesh process are the following:
The U and V dimension of the rectangular grid element used by the View Mesh.
If the Sample Size is small, this number will control the sample size so that an exorbitant amount of memory is not used.
When the selected visible cloud points are Z-buffered into the View Mesh grid structure, the Z-buffer tolerance determines whether new points replace existing points or whether they are averaged with existing points that are within the given tolerance of the given point.
This number is set based on the expected noise levels of the scanner.
Once the View Mesh is populated with all the visible selected cloud points, unoccupied elements of the grid are filled in by checking for defined elements of the grid within the given "Fill Size". The unoccupied elements are filled in by linearly interpolating the horizontal and vertical neighbors in the grid that are already filled in.
Other considerations when making a view mesh include:
You can save your current view settings as either a Named View or as a Temporary View using the TmpV tab. Saving a view is helpful in a variety of different situations. In the case of view meshes, saving a view allows you to document and redo the Make View Mesh procedure later.
EvalViewer allows objects to be either visible or invisible, and either selected or not selected. All objects are visible and selected by default.
You can subset cloud data before converting it to a view mesh to control what gets meshed. Only visible points are meshed.
The shaded View Mesh is a polygonized surface organized as a regular grid structure that allows for undefined (unoccupied) elements.
A View Mesh is exactly mathematically equivalent to a trimmed degree 1 (non-rational) B-Spline surface and could easily be converted to such an entity. However, very few systems can deal with B-spline surfaces with 500x500 control point grids.
Use the Smooth button to smooth a View Mesh any number of times, although the geometry of the shape will change slightly with each smoothing operation. Surface Normals are recomputed during each smooth operation.
All view mesh objects that are visible and selected are used as input.
The parameter relevant to the Smooth View Mesh function is the Smooth Limit, accessible from the Mesh > Params button, which limits the amount of smoothing that can happen to a mesh. It should be set to a value that is as large as the amount of noise that is present, but small enough to ensure that no modifications are too large.
All view mesh objects that are visible and selected are modified with a single 3x3 binomial smoothing filter. Each point location could be modified slightly. Each quadrilateral polygon normal is also modified slightly. The surface appears smoother.
The CurvMap operator computes curvature on the view mesh grid structure. Finite difference methods are used to estimate derivatives, and derivative estimates are used in standard curvature formulas to estimate curvatures. The default curvature map is the Mean Curvature map. Other curvatures are selected using the Surf tab. In the current version, you must also use the Surf > None button to turn the color curvature map off.
All View Mesh Objects that are visible and selected are used as input.
All View Mesh Objects that are visible and selected are displayed as color-mapped curvature objects. You can adjust the color bar range using the Incr and Decr buttons or the Up and Down arrow keys.
The Boundary button creates line boundary curves from View Mesh Objects.
All View Mesh Objects that are visible and selected are used as input.
A boundary line set is computed for each View Mesh object that is visible and selected.
You can define your own view mesh boundaries using the subsetting/cropping operators in the Subset tab with the Boundary button. Once boundary lines are computed, you can use all the operators on the Line tab to edit/modify/filter the shape of the boundary curves.
The number of polygons in a view mesh is usually large! To query the number of polygons and the number of points in a view mesh, choose the menu item Surface > Surface Info.
The polygon mesh derived from cloud data can often be directly used for various applications. The Reduce option allows you to reduce the number of polygons while remaining faithful to the original geometry by using a chordal deviation test.
The Mesh Reduction Tolerance in the dialog box invoked from the Params button. All view mesh objects that are visible and selected are used as input.
The output of the reduce operator is a new topological arrangement of quadrilateral polygons. To see it, use the tab button Display > PgnOutlines. Note that all the polygons and all the points are still present after the Reduce button is applied. However, some of the quads and some of the points are hidden.
Use the Reset button to go back to the original densely-spaced quad mesh topology. For example, if you smoothed the mesh but did not get as much reduction as you might like, you can Reset the topology, smooth some more and then reduce again until you are happy with the results. The Reset button does not reset the actual point locations.
Once you are happy with the results, use the Compress button to get rid of any internal details that are not being drawn or used. You will notice that if you choose the menu Surface > Surface Info, the number of polygons and points is drastically reduced. Typically, you can get rid of 90% of the original data and the surface will still look very good. Moreover, any XYZ sections that are cut through the surface will be very light with points.
The Erode button blanks out polygons near the boundary of a view mesh.
It is not unusual to have the surface represented by a point cloud turn away from your view near the boundaries of a view mesh. Imagine how the earth's surface at the equator looks if you are directly over the north pole. If this part of the surface is not suitable for further processing, you can strip it away by "eroding" the edge of the mesh.
All view mesh objects that are visible and selected are used as input.
Each mesh is maintained on output to be identical to the input mesh except that one layer of "sample size" polygons has been stripped away.
The Dilate button is the logical opposite of erode, and is used to expand a view mesh to fill small holes. Try this button if you do not get the view mesh results that you expect.
All view mesh objects that are visible and selected are used as input.
Each mesh is maintained on output to be identical to the input mesh except that one layer of "sample size" polygons has been added.
Consider ideal cloud data from a sphere. This data can be categorized as belonging to the front hemisphere that is visible, and the back hemisphere that is not visible. In general, you should use the Cloud tab operators to subset the data points into these two hemispheres before making a mesh.
However, in typical processing it is sometimes difficult to prevent the cloud data from front and back surfaces from "mixing together" in the mesh. The M-Filter (Maximum Filter) button looks in small neighborhoods and tries to make sure that only points associated with the front surface are used in the view mesh.
The Max Filter is not fool-proof. Sometimes, multiple applications of the filter can improve the results.
The M-Filter Size in the dialog box invoked from the Params button sets the distance between the front cloud and the back cloud surface.
All View Mesh Objects that are visible and selected are used as input.
Better looking surfaces with less mixing between the front cloud and back cloud (if present).
Blanks out polygons with normals orthogonal to the sight line used in view mesh creation.
The Normal Filter provides another criterion for getting rid of polygons that do not belong in a high quality model. Each polygon's normal is tested against the sight line and if it falls with a certain number of degrees orthogonal to the sight line, the polygon is blanked.
The N-Filter Angle threshold in the dialog box invoked from the Params button.
All view mesh objects that are visible and selected are used as input.
Each mesh is identical on output to the input mesh, except that polygons with normals orthogonal to the Z-direction of the view mesh are removed.
The S-Filter button blanks out polygons with long sides.
The size filter provides yet another criteria for getting rid of polygons that do not belong in a high quality model. This filter uses a criterion that all polygons must have a side length that is smaller than a given threshold.
The S-Filter Length threshold in the dialog box invoked from the Params button.
All View Mesh Objects that are visible and selected are used as input.
Each mesh is identical on output to the input mesh, except that polygons with sides longer than the given maximum length are removed.
The dialog box invoked from the Params button sets the parameters used by the following view mesh operators:
The U and V (horizontal and vertical) dimension of the rectangular grid element used by the View Mesh.
Sets the maximum number of samples (polygons) per U,V direction that EvalViewer will generate.
The smaller the sample size, the more polygons EvalViewer generates. If the sample size is so small that this value would be exceeded, EvalViewer adjusts the sample size to generate only the maximum number of samples.
Limits the amount of smoothing that can happen to a mesh.
Sets the minimum distance between the front cloud and the back cloud surfaces.
Each polygon's normal is tested against the sight line and if it falls with a certain number of degrees orthogonal to the sight line (the N-Filter Angle), the polygon is removed.
Polygons with sides longer than the given maximum length are removed.
Determines whether new points replace existing points or whether they are averaged with existing points that are within the given tolerance of the given point. Used during View Mesh creation.
Limits the size of holes that are filled in during the view mesh creation process. Pixels and grid elements are synonymous.
The chordal deviation tolerance used to decide whether a collection of 4 quadrangles in the view mesh is really just a single flat quad.
The Reset button resets a View Mesh to an "almost-like-the-original" View mesh if it has not been compressed. It undoes the topological changes done by "Reduce."
All View Mesh Objects that are visible and selected are used as input.
Resets the polygon topology to the state it was in after the mesh was created. Minor point movements within the Mesh Reduction tolerance are not reset.
The Compress button compresses a view mesh to the actual number of vertices and polygons needed to represent the shape. You cannot reset a view mesh once it has been compressed.
All View Mesh Objects that are visible and selected are used as input.
Polygonal objects with quadrilateral polygons.
![]() | Next: Temporary View tab |