|
|
Home : EvalViewer : Appendix :
 | Previous: Net Session testing |
This section contains answers to a number of frequently asked questions regarding the Cloud Data functionality in both EvalViewer and the Point Cloud toolset in Studio/Autostudio/SurfaceStudio/Power Animator.
Point Cloud Data is 3D digitized data defining a part or object. It is obtained with some type of measurement device.
Point Cloud Data can be obtained in a number of different ways, using a number of different technologies. In general, data can be obtained using two methods:
In the contact method, a device such as a touch probe; Faro, Romer, or Immersion arm, or coordinate measuring machine is placed on or travels across the object, stopping at predefined points along a set of axes to take an X, Y, Z measurement.
The non-contact method uses devices such as line or scanning point lasers, and photogrammetric systems. A type of light "beam" is reflected back to the measuring device, and through the laws of trigonometry the X, Y, Z point in space is determined.
This data is useful to the designer or engineer using Studio/Autostudio/SurfaceStudio who needs to obtain meaningful digital information from an existing part or object when no such information may exist. For example, a sculptor may have created a model head for an action figure. This model was created by manual/analog means and no digital database for this form exists. For accurate duplication and reproduction of this part a database of this form will need to be obtained.
A shoe designer may need to have existing foot forms or lasts digitized into surface data so that this information can be shared among many design/development and engineering locations. A digital file can be transmitted from location to location and between CAD technologies much more easily than a physical model. In addition, only one master of an object may exist, and if it is damaged it will have to be regenerated at a substantial cost.
In another situation, a clay modeler may have made some changes to an existing car model for a design "facelift". This may take the form of adding some details to an existing part/object from which this data will have to be extracted in order to be manufactured.
In all of these situations and more, having the cloud point data of the object will be the "starting point" to reverse-engineer a database and/or a manufacturable part from the master analog model.
The point cloud functionality within Studio/AutoStudio/SurfaceStudio currently supports 4 cloud data formats. These include Cyberware, and Hymarc, IGES (as entity 106, type 2, copious data) as well as the Alias wire file.
The stand-alone EvalViewer supports Cyberware, EOIS, Hymarc, Kreon, Sharnoa, Steinbichler, IGES (106/116), ASCII (X,Y,Z), VDA (P Set), and the Alias wire file as well as the EV-Geo (EvalViewer Geometry) file.
With the ability to move clouds back and forth to Studio using the clipboard's copy and paste functions, the workflow between EvalViewer and Studio is greatly enhanced. The new EV-Geo file replaces the EV-Cloud and EV-Line files since it holds cloud, line, curve, polygon, and surface data, including view meshes.
EvalViewer currently exports IGES and ASCII XYZ as cloud data file output formats as well as Alias Wire and EV-Geo files. EvalViewer has the ability to export line section data as IGES, DES, DXF or Catia-CAI as well as Alias Wire or EV-Geo. EvalViewer can export polygonal data as either:
No, presently a cloud data file is displayed within the file lister like any other binary format file. In EvalViewer, there is a single File > Open option. The multiple file selector reads the various cloud data file formats available and is able to discern between one vendor file and another.
Currently you will have to actually attempt to load the file into EvalViewer or Studio/AutoStudio in order to confirm that you actually have cloud data. However, some vendors apply an extension to their actual cloud data file which may aid in identifying the file as an actual cloud data set. Some extensions are:
This is not a recommended practice at the present time for a number of reasons. It may be logical that if a dataset can be converted to polygons and exported to Studio as an.obj format file, then an STL format file could subsequently be output. However, the quality of the polymeshes from Clouds to Polygons that EvalViewer outputs are not STL quality. This means that there may be gaps, holes and improperly orientated polygons, all of which would cause problems for an RP process accepting this low quality file.
EvalViewer supports view meshes. View meshes are "water-tight" and single-valued except at visible holes unlike, the "Clouds to Polygons" meshes. After compression, some rapid prototyping systems may accept the View mesh, but you will have to build the enclosing surfaces because EvalViewer does not automatically create a single mesh from multiple view meshes.
Tools within the Alias Systems family that support cloud data include the stand-alone EvalViewer and the Point Cloud/Cloud Fit functionality within Studio/AutoStudio/SurfaceStudio.
It is possible to create surfaces directly from cloud data using the Point Cloud tools within Studio/AutoStudio/Surface Studio. The most robust means of creating elegant (Class A) surfaces is through a process of sectioning the data within EvalViewer, exporting this section data as DES to Studio/AutoStudio/SurfaceStudio, and then creating/fitting spline geometry in Alias with this data. In EvalViewer, a tool allows you to test-fit the cloud section data (before exporting) by building a "quick and dirty" surface over it.
In EvalViewer, there is a Cloud > Surface Fit option that can be used for evaluating the quality of cloud data and for very simple surface fits of smooth slab surfaces. The surface that is created can be copied to the clipboard and pasted into Studio.
The amount of data that can be handled by EvalViewer is limited by the amount of RAM your machine has. One rule of thumb to use is this:
Maximum Number of Points = 83,000 x (Amount of RAM in MB on Your Computer - 70)
Total Cloud Data File Size = (Amount of RAM in MB on Your Computer - 70) megabytes.
For example, if you have 128 MB of RAM, then 128-70 =58 megabytes of Cloud data can be processed, which would be about 4.8 million points.
The actual minimum amount of RAM that you need for your data is 12 times the number of data points that you want to process. If you want to process a cloud data file containing 1 million points, you need 12 MB of RAM above the normal system overhead requirements at the time.
Generally, increasing swap only makes things run very slowly. You can usually tailor your work process to avoid overflowing your real memory. When you are processing cloud data, every visible point is examined on each pass through the data. When swapping begins, it is usually faster to kill the job and start over, adopting a new data processing sequence that prevents so much data from being in memory at once. In other words, you can process two halves of the job at full speed much more quickly than if you try to process the whole job at once and your machine starts swapping.
EvalViewer was designed for, and is best suited at present as, a cloud data pre-processor. It can filter, smooth, and reduce extraneous data by both automatic and manual means. This allows you to eliminate noisy and unwanted data rapidly and easily. You can then use this "optimized" data set as a reference for section lines and the determination of character lines for modeling. The section data can then be exported from EvalViewer and imported to Studio/AutoStudio/Surface Studio for subsequent surface development.
In these packages, the Point Cloud/Cloud Fit functionality was designed to fit surfaces to cloud data sets. This functionality is complementary to EvalViewer when EvalViewer is used as the pre-processor.
The file format best suited for data exchange between the two applications is the Alias wire file itself. All data from EvalViewer can be sent to Studio using the wire file and almost all geometry from Studio can be sent to EvalViewer via the wire file. Moreover, EvalViewer now supports copy (Ctrl-C) and paste (Ctrl-V) to/from the clipboard. Cloud data can also be transferred via IGES. For section cuts, the DES or IGES options may also be used.
One caveat is that EvalViewer sends lines to the Alias wire file as degree-1 splines. However, degree-1 splines from Studio are now automatically converted to lines when imported into EvalViewer. This minimizes the number of points on the lines, makes it possible to use EvalViewer's line tools, and improves interactive response.
Yes, if you output your sections as IGES, that data should go into almost any other CAD system. If you want to import surface data into EvalViewer, you can export IGES or VDA from your CAD system. If you have SDRC, an SDRC plug-in option is available from Alias Systems to convert directly to an Alias wire file, which EvalViewer will read. If you have CATIA, the EvalViewer CAI interface also provides utilities to convert CATIA data to the Alias wire file format.
EvalViewer is not yet a "push-a-button-and-the-data-is-processed-and-a-model-gets-created technology." EvalViewer is a focused tool for pre-processing a wide range of cloud data vendor formats from 3D laser scanners. It is used to create cross section and feature curve data for import to Alias for subsequent surface development. You will have to develop your own practices based on experience with your particular datasets and an understanding of the tolerances within which they must work. You should have an understanding of the tools and practices recommended for processing the data as well as experience with the suite of surfacing tools provided within Studio/AutoStudio/SurfaceStudio.
The key advantages of cloud data are that it can be digitized relatively quickly and that it can almost completely capture the outer surfaces of any object. The ability to easily visualize the cloud data offers you a way to keep the object (for measurements and visual comparisons) no matter where the physical part goes. Conventional CMM-touch probe data cannot provide this option.
In regard to the Point Cloud functionality, you should know what is possible and reasonable to accomplish with the cloud surfacing tools in order to achieve expected results.
One group of functions known as Cloud Fit is integral to EvalViewer. This group of functions has the following properties:
Another group of functions is in the EvalViewer program. This group of functions has the following properties:
Section data must be exported from EvalViewer as either IGES or DES data. You can send data from Sharnoa, EOIS, Kreon, Steinbichler, and VDA through EvalViewer to get it into an IGES cloud format suitable for Studio.
By packaging Cloud functions with the EvalViewer stand alone program, you can interactively process over a half a million more points in a given job than you could do otherwise (given that your computer only has a fixed amount of RAM).
Some cloud formats do not have sufficient header information to seamlessly integrate them into StudioTools' data browser. Therefore, only some vendors were selected for the direct Studio Open/Retrieve mechanism. The choice was also based on customer requests and available resources.
Direct surfacing of large datasets is something that is being discussed for future versions. The stand-alone currently provides one surface generation function under Clouds > Surface Fit, as well as Skin and Square tools (in the Polygons menu) that build quick NURBS surfaces over a set of lines to test-fit cloud data sections before exporting to a modeling package. Surfaces can also be modified and written back out to Alias wire files.
The most-used cloud functions in EvalViewer, such as Quick-Render Cloud and Section Cuts > Multiple Clouds, are based on the assumption that the data points are acquired in what one might call "optical scanner order."
For example, most Hymarc files, all Cyberware files, all Steinbichler.ac files, all EOIS.xyz files, most Kreon files, and most Sharnoa files are compatible with this assumption. However, not all cloud data files are compatible. The following documents a few of the limitations that we have encountered.
Hymarc users should be aware that the Hymarc scanner can be run in a forward mode, a backward mode, and a forward-backward mode. EvalViewer is designed for cloud data from the Hymarc scanner that is acquired in either the forward mode or the backward mode. It will not quick-render or section clouds correctly with cloud data that is acquired in the forward-backward mode. This applies also to Sharnoa scanners that may acquire data in a CNC lace cut or snake mode.
You have the option of using view meshes or voxel meshes for data that won't quick-render. View meshes work well for processing any cloud data as long as the view-dependent aspects are not a burden, while voxel meshes do not depend on viewing direction. Both view meshes and voxel meshes are completely independent of point order so that forward-backward Hymarc data is not a problem.
If you preprocess most types of Cloud data in Imageware and then write a thinned-out file, these points will not be the same as when they come directly from the scanner. EvalViewer prefers point clouds directly from the optical scanners.
Owing to the number of service bureaus that already have Imageware, it is likely that you may receive an IGES file from a service organization. Now that you have EvalViewer, you should explicitly request that you want the raw data in the scanner format if they use one of the scanners that we support.
In addition, view meshes and voxel meshes are not dependent on point ordering so these methods of cloud processing are a good option even when the data does not come directly from the scanner.
The PixSys scanner does not create "optical scanner data." It is similar to data from touch-probe scanners like the Faro, the Romer, or the Immersion. If you acquire data with a touch-probe either in a section format or a surface boundary format, you will be able to process the data directly in your surfacing package. If you are going to acquire dense data with a touch probe for use in EvalViewer, you should stick with a right-left, right-left ordering (or a left-right, left-right ordering), but do not do random scribble lines or alternate right-left, left-right.
If you use view meshes or voxel meshes in EvalViewer, you can acquire Pixsys data in any order you want. It is not sensitive to point ordering.
Data from some computed tomography systems does not come in optical scanner order. It comes in a "closed contour" "constant depth" format. EvalViewer can convert pure cloud data from CT scanners into lines (using the Clouds > Cloud to Lines tool) that can then be smoothed and data reduced.
However, Cloud QuickRender does not usually work, and therefore Multiple Cloud-Cloud sections does not work well either.
View meshes are not as easy to use with CT data as they are with optical scanner data, but they can be used. You must use the cloud cropping tools to crop out the appropriate points prior to meshing. However, do not confirm before view meshing (it is not necessary, and you will have to read in a fresh copy of the file each time).
If you start drawing a line cut and you do not want to complete it, complete the drawing of the line and make sure it is very short (that is, less than 10 pixels). This will cause the system to ignore it. When screen overlay entities are drawn that are smaller than 10 pixels, no operation is done.
 | Next: Supported scanner vendors |