johng@OCE.ORST.EDU (John A. Gregor) (05/30/91)
In article <6154@ns-mx.uiowa.edu> randy@tessa.iaf.uiowa.edu (randy frank) writes: > I have data for a variable sampled at random points within a given > area. I would like to build an image from this data but this requires > resampling the data on a fixed grid. This sounds very similar to a request I posted a couple of months back. And since I never got around to summarizing and thanking the folks who responded, I figgured that this would be a perfect opportunity. First, what I wrote: Me> We have a relatively large (several thousand) set of ocean samples in Me> 3-D that we want to visualize. Unfortunately, all the tools we have Me> demand that the data be represented as gridded data or as a connected Me> mesh (either tetrahedrons xor hexahedrons). Any pointers to routines Me> or literature to accomplish this would be greatly appreciated. Me> Some more notes about the data: It's scattered in XY but linear in Z. Me> Also, the number of and distance between Z samples varies among the XY Me> locations. Second, I would like to thank the following folks for taking the time to respond: * - wes@ux5.lbl.gov (Wes Bethel) ^ - nwatson@enuxha.eas.asu.edu (Nathan F. Watson) * - uselton@wk207.nas.nasa.gov (Samuel P. Uselton) * - "John D. McCalpin" <mccalpin@perelandra.cms.udel.edu> * - sher@BBN.COM ? - apas611@hermes.chpc.utexas.edu (David Boles) ? - Al Boehnlein <ajb@iti.org> ^ - Peter Hughes <pete@saturn.ucsc.edu> ^ - lane@enuxha.eas.asu.edu (David A. Lane) ? - Tegenf{lt <tegen@isy.liu.se> ? - ianh@bhpmrl.oz.au (Ian Hoyle) * - Shai Guday <shaig@shum.huji.ac.il> * - Andrea Hin <hin%svcentld%HDEDH1.wldelft.nl@CUNYVM.CUNY.EDU> legend: * - pointers to products, papers, techniques and other resources ^ - pointers to people who are doing similar things ? - They asked for information that I found The request I posted was for a colleague. He chose to write his own routine to generate an interpolation function. He then just sampled at regular X, Y, and Zs. So we never did construct a mesh of tetrahedrons from the scattered points. -- John A. Gregor College of Oceanography E-mail: johng@oce.orst.edu Oregon State University Voice #: +1 503 737-3022 Oceanography Admin Bldg. #104 Fax #: +1 503 737-2064 Corvallis, OR 97331-5503 ######################################################################## From: wes@ux5.lbl.gov (Wes Bethel) I have an avs module which does this. The underlying number cruncher is based on a quadratic sherard (sp?) approximation method. This source code is available from netlib. If you've never used netlib before, check it out. First, send mail to netlib@ornl.gov. The one-line body of the message should be "send index". The category you're interested in is TOMS (Transactions on Mathematical SOftware). So, you're next email to netlib will be "send index from TOMS", though I can't remember if case sensitivity is employed. The algorithm number is something like 661. Grep for "scattered" on the toms index and you'll find what you're looking for. Oh yeah, I hope you have a fortran compiler. Good luck, wes [ I would strongly recommend that everyone take a look at what netlib has. Also, I'd recommend sending the first part of the second paragraph to anyone asking for anything in the graphics or visualization newgroups. -JG ] ######################################################################## From: nwatson@enuxha.eas.asu.edu (Nathan F. Watson) John Gregor, I am a student at Arizona State University. I hope you find a suitable algorithm to visualize your data. I'd just like to let you know that a friend of mine is working on the very problem you describe (i.e., visualizing data sampled in random (x,y) locations at varying z depths). Though your need may be short-term, you might profit by contacting him, and I am sure he would appreciate more real-world data. I believe you can contact him at e-mail address lane@enuxha.eas.asu.edu, and if not, you can always mail me. You can reach him by USMail at David Lane Computer Science Department Arizona State University Tempe, AZ 85287-5406 The CAGD faculty at Arizona State University are committed to research in scientific visualization and have several recent publications dealing with data interpolation and visualization. Thank you very much for your attention. Nathan F. Watson nwatson@enuxha.eas.asu.edu [ David Lane also sends mail later -JG ] ######################################################################## From: uselton@wk207.nas.nasa.gov (Samuel P. Uselton) Fluid dynamicists are starting to use "unstructured" grids for computational simulations. Points are distributed to put higher density in "more interesting" regions. Programs are used to make tetrahedra based on these points with highly variable spacing. Typical grids have HUNDREDS of thousands of points. They basically start from something like the Voronoi diagram's dual (see also Delaunay, Dirichlet tesselations). Best sources for algorithm desciptions: books on Computational Geometry, eg Edelsbruner or Preparata & Shamos. There may be software floating about, but I don't have it. hope this helps. Sam Uselton uselton@nas.nasa.gov employed by CSC working for NASA (Ames) speaking for myself ######################################################################## From: "John D. McCalpin" <mccalpin@perelandra.cms.udel.edu> For synoptic data, I would recommend a scatter-plot technique in three dimensions. I would use hue to represent the amplitude of the field (salinity or temperature) and using depth-cuing (varying 'value' in the hue-value-saturation system) to represent the distance from the viewer. This is pretty easy to do on a Silicon Graphics machine. The hardware limitation is that you have to divide the data into a fairly small number of hues because you need to fit the color ramps for each hue into the color table. -- John D. McCalpin mccalpin@perelandra.cms.udel.edu Assistant Professor mccalpin@brahms.udel.edu College of Marine Studies, U. Del. J.MCCALPIN/OMNET [ Yes, scatter plots are another useful visualization technique. However we were specifically trying to create contour plots and isosurfaces, which require grids (or some other interpolation -JG ] ######################################################################## From: sher@BBN.COM Using our true-3-D SpaceGraph display, we can simply display the points as points, wherever they are. If you have any opportunity to visit here (BBN, Cambridge, MA), I'd be happy to show you, with your data. -- .---------------------------------------------------------------- /Internet email: sher@bbn.com Larry Sher < US Mail: BBN, MS 6/5A, 10 Moulton St., Cambridge, MA 02138 \ Telephone: (617) 873 3426 FAX: (617) 873 3776 `---------------------------------------------------------------- ######################################################################## From: apas611@hermes.chpc.utexas.edu (David Boles) Hello John, I work at a sonar research lab and would be very interested in the results of your request for information about data visualization. If it isn't too much trouble, could you please email me the replies that you get. Have you looked at volumetric rendering packages? David Boles Applied Research Labs University of Texas at Austin apas611@hermes.chpc.utexas.edu [ Yes, we have AVS on a DECstation 5000-200, apE, Wavefront Data-Visualizer, Wavefront Advanced Visualizer + homebrew code, I think we have slicer on some macs, etc. Lots of programs, dozens of file formats, never enough disk space, etc. -JG ] ######################################################################## From: Al Boehnlein <ajb@iti.org> I'd be intrested also in the replies that you get. I have a simular problem with Z data from a gaging instrument. ajb ######################################################################## From: Peter Hughes <pete@saturn.ucsc.edu> A person I know who has worked with similar oceanographic data is Cindy Ferguson (cindy@spica.ucsc.edu). ######################################################################## From: lane@enuxha.eas.asu.edu (David A. Lane) Dear Dr. Gregor, [ Sigh, not yet. 'John' will do nicely :-) -JG ] I have read your article in comp.graphics.visualization regarding visualization of a large set of ocean samples in 3D. One of the topics in my dissertation, which I am currently working on, is the interpolation and visualization of well-log data. From your description of the data, I believe your data are similar to well-log data, i.e. data sampled densely at arbitrarily located wells. Currently, I am working for Dr. Thomas A. Foley at Arizona State University, and we are developing a method for interpolating well-log data. We have tested our method on several set of mining data and if you are interested, we can apply our method to your data. We also have developed a program to visualize well-log data. We look forward to hear from you. David Lane Computer Science and Engineering Department Arizona State University Tempe, AZ lane@enuxha.eas.asu.edu (602) 965-3245 =========================================================================== From: Tegenf{lt <tegen@isy.liu.se> In comp.graphics.visualization you write: I don't have anything you might have any use for, but I would like to hear from you if/when you've got some replies on your posting. My main interests are divided into two areas, fast photorealistic raytracing and visualization of 3D volumes of different kinds. Yours sincerely Clas Tegenfeldt Linkoping University Sweden tegen@isy.liu.se =========================================================================== From: ianh@bhpmrl.oz.au (Ian Hoyle) John, I'd *really* appreciate it if you could pass on any responses that you get. I have a not so dissimilar problem using reservoir simulation data which has the data points lying on what I would loosely call a spatially deformed 3-grid :-) ie. the grid follows the geology so the voxel elements are irregular hexahedrons. I've been playing with Wavefront's new data visualizer on the problem with some success, but it is early days as far as the development of that product goes. My machine is a SiliconGraphics 240VGX and what I*would* like to try out is using VoxelView (which is REALLY fast on our VGX) on the data regridded to a uniform 3-D grid. Anyway .... I hope you turn something up !! ian -- Ian Hoyle /\/\ Image Processing & Data Analysis Group / / /\ BHP Research - Melbourne Laboratories / / / \ 245 Wellington Rd, Mulgrave, 3170 / / / /\ \ AUSTRALIA \ \/ / / / \ / / / Phone : +61-3-560-7066 \/\/\/ FAX : +61-3-561-6709 E-mail : ianh@bhpmrl.oz.au [ It sounds like you have an easier problem, you know the topology of the mesh already. Then, for each point in your uniform grid, determine what hexahedron you are in and interpolate (or average, or whatever). Good luck! -JG ] =========================================================================== From: Shai Guday <shaig@shum.huji.ac.il> You neglected to say what your graphical system is. I have been searching for such a system that is compatible with X11R4. The closest thing that I have found is gnuplot3d - which is compatible with several other systems as well, though I am uncertain as to its capabilities in handling such a large number of data. I have recently been sent to the source code of a program called xwave which is used to project waves upon a mesh. This code (~350 lines) is relatively simple and I will be using it as the basis of a larger program with animation capabilities. If you do recieve other answers or suggestions, please let me know as I would be most interested in saving the hours I plan to spend on the project. -- Shai Guday This product is all-natural and MOSIX System Group contains NO artificial colorings Hebrew University or artificial preservatives. email: shaig@hujinix.huji.ac.il [ Graphics system depends on the tool. Some use X, some sunview, some GL (SGI & IBM), some starbase (HP), some PHIGS, some PEX, some quickdraw, some postscript, some tek 4100, etc. The common feature is that most only talk to gridded data. -JG ] =========================================================================== From: Andrea Hin <hin%svcentld%HDEDH1.wldelft.nl@CUNYVM.CUNY.EDU> Dear John Gregor, I read your mail about visualizing irregular data. Try to contact Dynamic Graphics, Inc. They have a program called the Interactive Volume Modeling System. It is able to read irregular data, although they transform it in a preprocessing step into a regular data set. Fun is that it can indicate the data points of the irregular grid by drawing vertical lines from the surface to the point of measurement. You can contact them: 1015, Atlantic Avenue Alameda, CA 94501 USA Phone 415-522 0700 Fax 415-522 5670 By the way, the IVM program is a commercial product! I have no details on prices. GrReEtEuTrInNeGdS, Andrea Hin Delft Hydraulics The Netherlands