[vtkusers] Interpolation of data..

Rakesh Patil rakeshthp at in.com
Mon Apr 26 13:03:20 EDT 2010


 Hi,As far as my knowledge is concerned, a point with unknown value can lye either inside a range [min, max] or outsie the range [min, max].If the point lies inside the range, then using interpolation we can find the unknow value for that point.. And if the point lies outside the range,then, using extrapolation technique, we find the unknown value for that point.. I guess I'm right.. If i'm wrong please tell me..Here in my case, i need to find the same elevation at lat/lons for all the node points on my mesh.. I guess inverse weighted method should do the need.I also read that shepard method uses inverse weighted method, but it doesnt suffice my need.. I hope my problem is clear.. What values it finds for the nodesinside the range, the same value i want to be found for the node points that fall outside the range..I've seen that some plottin tools have the capability to do this... But i'm blank about how that is done.. I have implemented the linear interpolationmethod and it w
 orks fine. It interpolates and gets the values for all points (inside and outside).. But it takes hell amount of time.. So, i want to overcome the speed.Any idea or suggestions..??ThanksRegardsRakesh Patil Original message From:Bill Lorensen< bill.lorensen at gmail.com >Date: 26 Apr 10 17:35:48Subject:Re: [vtkusers] Interpolation of data..To: Rakesh Patil What values would you expect in area where you have no elevation data?Maybe I misunderstand the problem.On Sun, Apr 25, 2010 at 9:45 PM, Rakesh Patilwrote:>>> Hi,>> So, you mean there is no other way..?? extrapolation or anything..??>> In that case i need to write my own filter to do this then..??>> Thanks>> Regards> Rakesh Patil>  Original message > From:Bill Lorensen< bill.lorensen at gmail.com >> Date: 26 Apr 10 02:13:14> Subject: Re: [vtkusers] Interpolation of data..> To: Rakesh Patil>> If you do not have data in a region, I don't see how you can interpolate it.>> On Sun, Apr 25, 2010 at 1:54 PM, Rakesh Patil wrote:>>>>>> Hi
  Bill,>>>> Thanks for your reply.. I implemented that technique and it works fine. As>> i>> said in my previous post, outsid range values are not interpolated.>>>> For example, pls see the attached image file for reference.. In the>> figure,>> the mesh which is generated, doesnt completely fall in the area of the>> scatter data..>> In the rwo white patches, the values are not interpolated. I mean i dont>> get>> the actual result there in those areas (as there is no scatter points>> avail>> there)>>>> So, what can i do to get values for my entire mesh instead of getitng for>> partial mesh.??Either interolation or extrapolation.. Any means is>> available>> in VTK>> to implement this..??>>>> Thanks>>>> Regards>> Rakesh Patil>>  Original message >> From:Bill Lorensen< bill.lorensen at gmail.com >>> Date: 24 Apr 10 18:15:33>> Subject: Re: [vtkusers] Interpolation of data..>> To: Rakesh Patil>>>> The last part of this example:>> http://www.vtk.org/Wiki/VTK/Examples/InterpolateTerrain
 >> does what you want.>>>> Basically,>> 1) Add the lat/lo g and elevations to a polydata>> 2) Create a surface with Delaunay2D>> 3) Add the points with unknown elevation to a vtkPoints>> 4) Use a cell locator to cast infinite rays into the surface to find>> the elevation.>>>>>> On Fri, Apr 23, 2010 at 9:50 PM, Rakesh Patil wrote:>>>>>>>>> Well,>>>>>> I was trying with this example:>>>>>> http://www.vtk.org/Wiki/VTK/Examples/InterpolateTerrain>>>>>> In this example, there are two ways specified for interpolation.. One is>>> using>>> vtkProbeFilter and another is using vtkCellLocator>>>>>> First i tried using vtkProbeFilter. But failed getting the output.. But>>> using the>>> vtkCellLocator, i was able to produce interpolated values for the third>>> column of MeshG.>>>>>> After observation, i foound that only the region which is common in both>>> the>>> grids,>>> was interpolated. So i was fully stuck wondering how to find the>>> remaining>>> values..>>>>>> any comments..??>>>
 >>> Thanks>>>>>> Rakesh Patil>>>>>>  Original message >>> From:"Dean Inglis"< dean.inglis at camris.ca >>>> Date: 23 Apr 10 23:59:48>>> Subject: Re: Re:[vtkusers]Interpolation of data..>>> To: "Rakesh Patil">>>>>> How can you expect your mesh to get interpolated>>> when its z coordinate values are all 0>>> but the scatter data range in z from 700 to 900?>>> The bounds of each unstructured grid do not even>>> overlap / intersect!>>>>>> Dean>>>>>>> Hi,My both the meshes are uns tructured grid. My application is based on>>>> FEA meth od.. ScatterG does contains scalars and the depth values are>>>> stored as a scalar array for scatterG..I dont want coordinate values to>>>> be>>>> interpolated.. As an example,scatterG contains data something like>>>> this,Points: 87.23 16.2690086.24 17.8753087.18 17.27700......and the>>>> scalar for scatterG is defined as 900530700......meshG contains>>>> points84.1217.83 081.3315.28 083.2418.28 0......So,, what i want to do>>>> is,>>>> using the th
 ird column of scatterG, i need to estimate the third column>>>> of>>>> meshG..Any ideas how it can be done.??ThanksRakesh Patil Original>>>> message>>>> From:"De an Inglis"< dean.inglis at camris.ca >Date: 23 Apr 10>>>> 20:16:47Subject:Re: Re:[vtkusers]Interpolation of data..To: "Rakesh>>>> Patil">>>& gt; vtkProbeFlter will probe the scalar values ateach spatial location>>> of>>>> your>>>> scattered points: youneed to have scalar, vector, or tensor data>>>> asso ciatedwith the grid's points:scatterG>GetPointData()>SetScalars(>>>> somevt>>>> kDataArray );the vtkProbeFilter will not interpolate any>>>> coordinatevalues>>>> like x , y,z locations.If you have depth valuesin a vtkDataArray that>>>> correspond to the x, y, z locationsin your scatterG grid, then this>>>> should>>>> work.Does anyone know ifmaybe the input to the probe filter needs to be>>>> a>>>> structured data set likevtkPolyData? By the sound of your grid's name>>>> "mesh" , this sounds like structureddata.Can it be
  converted to a>>>> vtkPolyData?Also, you allocate a smart pointer as:vtkSmartPointer probe>>>> =>>>> vtkSmartPointer: :New()Dean Original Message From: "Rakesh Patil" To:>>>> Cc:>>>> Sent: Friday, April 23, 2010 9:40 AMSubject: Re: Re:[vtkusers]>>>> Interpolation of data..> Hi..I can independently view those grids...>>>> This>>>> is what I have done..vtkUnstructuredGrid *scatterG =>>>> MyApp::getInstance()>getScatterGrid();vtkUnstructuredGrid *meshG =>>>> MyApp::getInstance()>getMeshGrid();vtkSmartPointer probe =>>>>>>>>>>>>>>>> vtkSmartPointer::New();probe>SetSource(scatterG);probe>SetInput(meshG);probe>Update();vtkDataArr>>>> ay *data = probe>GetOutput()>GetPointData()>GetScalar();vtkDoubleArray>>>> *doubleArray =>>>>>>>>>>>>>>>> vtkDoubleArray::SafeDownCast(data);MyContour::getInstance()>setDisplayArray(doubleArray);After>>>> doing this, i click on contour display button, and the contour for>>>> bathymetry gets displayed for meshG .Initially, meshG doesnt have>>>> depth/
 elevation (z) value. z value is always 0 for every node point. x>>>> and>>>> y corresponds to the lon and lat respectively.But the scatter data>>>> contains lon, lat, and depth.. So, once the interpolation is done, i>>>> should get the bathymetry (depth/elevation) for meshG also..'MyContour'>>>> in>>>> the above code is the class which displays the contours for wh> atever>>>> data is being passed.So, any suggestions or > comments..??ThanksRegards>>>> Rakesh Patil Original messageFrom:Dean Inglis< > dean.inglis at camris.ca >>>>> Date: 23 Apr 10 16:49:50 Subject:Re: > Re:[vtkusers]Interpolation of>>>> data.. To: "Rakesh Patil">>>>>>>>>>>>>> >>> Powered by www.kitware.com>>>>>> Visit other Kitware opensource projects at>>> http://www.kitware.com/opensource/opensource.html>>>>>> Please keep messages ontopic and check the VTK FAQ at:>>> htt p://www.vtk.org/Wiki/VTKFAQ>>>>>> Follow this link to subscribe/unsubscribe:>>> http://www.vtk.org/mailman/listinfo/vtkusers>>>>>>>>> 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20100426/f1d6f968/attachment.htm>


More information about the vtkusers mailing list