FOVEX 360s/Point Cloud Generation

From FOVEX Help
Jump to: navigation, search


Generating Point Cloud

Note: The technology of the point cloud generation with FOVEX 360S is multi-stereo image matching.


Note: Before the point cloud generation:
  • The images of the project should have been oriented.
  • The camera should have been calibrated with a good level of accuracy.


Clicking the following icon from “Auto Process” toolbar starts point cloud generation:

Generating Point Cloud


The following parameters at the Settings of the Project will have influence of speed and quality of point cloud:

The following parameters at the Settings of the Project will have influence of speed and quality of point cloud:


Parameter Value 
Dense Matching with emphasize Options for dense matching with:
  • Fill region but not clean edge
  • Clean edge but not necessarily fill region
Image resolution reduction by facto Options for image resolution reduction in dense matching:
  • 1 (Full size)
  • 2 (Each side factor 1.4)
  • 3 (Each side factor 1.7)
  • 4 (Each side factor 2.0)
  • 8 (Each side factor 2.8)
  • 16 (Each side factor 4.0)
Max. acceptable range (m) Points with range (distance to the camera stations) more than this value will not be included in the Point Cloud.
Use of single stereo pair
  • If not selected, point cloud of multi-stereo pairs are exported. This means that point with point cloud are computed with multi-ray intersection.
  • If checked, point cloud of single stereo pairs are also exported therefore the point cloud will include points with two ray intersection.


point cloud

The project before dense matching for point cloud generation. If the number of points is more than 50 million points, the Network View will not visualize the point cloud. You can use free third party viewer for the visualization for example Quick Terrain Reader.


The produced point cloud with 82 Million points.


File Structure of the Results

All results are stored in the “PointClouds” Folder which is created beside the FOVEX Measure3D project file. Per each run for point cloud generation a folder with the date and time of runtime is created and the results of that specific run is stored in the date and time folder. This folder will contain the following information:


Files Description
Log.txt The log of point cloud generation including the parameter settings
#.las The 3D colored point cloud in .las format
#.ply The 3D colored point cloud including normal vector per each point in .ply format


Note:
  • .las format can be opened for viewing by laser scanner viewing software. Quick Terrain Reader .is a free software viewer to view the colored point cloud.
  • .ply format can be opened for viewing with many third party point cloud processing software. .ply can contain normal vector data, mesh in addition to only colored point coordinates.

Stereo view

If you have anaglyph glasses (cyan-red or red-blue) you can see the objects of Network View in 3D by setting the view in anaglyph mode as shown in the following:

Stereo view


Anaglyph view of the Network View with a right-red glasses.

Point Cloud Filtering

The following two filters are available in the Point Cloud module of FOVEX 360s:

  • Noise reduction
  • Outlier removal

Noise Reduction

The noise reduction filter can be applied on the point-cloud by right click on the name of the point cloud at Project Explorer and pressing Noise Reduction:

Noise Reduction


The following dialog box will be appeared:

Noise Reduction-The following dialog box will be appeared


In which the search radius for the neighboring points should be set. By clicking on OK the process of noise reduction will start. This process is single threaded and may take long time for large datasets.

Outlier removal

The outlier removal filter can be applied on the point-cloud by right click on the name of the point cloud at Project Explorer and pressing Outlier Removal:

Outlier removal


The following dialog will be appeared in which two parameters has to be set:

Outlier removal-The following dialog will be appeared in which two parameters has to be set


Parameters Description
Number of nearest neighbors to use for mean distance estimation The number of points to use for means distance estimation.
Standard deviation multiplier for the distance threshold calculation The distance threshold will be equal to: mean + stddev_mult * stddev. Points will be classified as inlier or outlier if their average neighbor distance is below or above this threshold respectively. What this means is that all points which have a distance larger than stddev_mult * stddev of the mean distance to the query point will be marked as outlier and removed.


By setting these parameters and clicking OK the process of outlier removal will start. This process is single threaded and may take long time for large datasets.

Triangulation

3D triangulation for a mesh generation and surface modeling can be performed on the point cloud by right click on the name of the point cloud at Project Explorer and pressing Triangulation:


3D triangulation


The following dialog box will be appeared:

3D triangulation-The following dialog box will be appeared:


Parameters Description
Multiplier of nearest neighbor distance They control the size of the neighborhood. The former defines how many neighbors are searched for, while the latter specifies the maximum acceptable distance for a point to be considered, relative to the distance of the nearest point (in order to adjust to changing densities). Typical values are 50-100 and 2.5-3 (or 1.5 for grids).
Maximum number of nearest neighbor
Search Radius (mm) It is practically the maximum edge length for every triangle. This has to be set by the user such that to allow for the biggest triangles that should be possible.
Min. angle of each triangle (radian) They are the minimum and maximum angles in each triangle. While the first is not guaranteed, the second is. Typical values are 10 and 120 degrees (in radians).
Max. angle of each triangle (radian)
Max. normal deviation (radian) They are meant to deal with the cases where there are sharp edges or corners and where two sides of a surface run very close to each other. To achieve this, points are not connected to the current point if their normal deviate more than the specified angle (note that most surface normal estimation methods produce smooth transitions between normal angles even at sharp edges). This angle is computed as the angle between the lines defined by the normal (disregarding the normal’s direction) if the normal-consistency-flag is not set, as not all normal estimation methods can guarantee consistently oriented normal. Typically, 45 degrees (in radians) and false works on most datasets.
Normal Consistency


By setting these parameters and clicking OK the process of triangulation will start. This process is single threaded and may take long time for large datasets. The following pictures shows the point cloud and the result of triangulation.


Point cloud.


Triangulated point cloud. This view shows only part of the triangulated mesh.

Point Cloud Difference

Two point clouds can be compared and the difference of the two point clouds can be displayed as a heat map. For point cloud difference computation, right click on the point cloud from Project Explorer that you would like to compute its difference to another point cloud and press Difference to …:

Point Cloud Difference


The following dialog box appears. Select the second point cloud (it will be the reference point cloud) and press OK:

Select the second point cloud (it will be the reference point cloud) and press OK


Once the process ends, the point cloud difference is added to the list of Point Clouds in the Project Explorer and will be saved on hard disk in the folder of “Point Cloud” (which is located beside the project file)

Each point cloud difference has its own legend value. Click on the point cloud difference at Project Explorer to show the legend as shown in the following figure:

Displaying the legend. The number in front of colors shows the difference value of point cloud. The red regions have the largest difference.