Point cloud for GIS use

    
  Subscribe Topic

Rank: 1

OldForumPost

Newbie

posts: 0

Registered: 2012-1-14

Message 1 of 13

 Point cloud for GIS use
14-09-2005 10:48 . am   |   View his/her posts only
One of my potential customers wants us to produce a surface of a the sea bed of a propsed coal terminal. This naturally is very large; the pier alone is 1500m into the sea. This will be used to design the piles, concrete decks and steel structure onto. They have a series of points for the surface. Has anyone done a job of this scale? Any hints on the best procedure?

ocp

Rank: 1

Steve

Newbie

posts: 0

Registered: 2004-3-1

Message 2 of 13

14-09-2005 02:21 . pm   |   View his/her posts only
It does not make sense to use a 1 to 1 scale for the data. There are limits to the view extents in VX that will not allow you to view data that is a mile from the origin. Also, you would have to adjust the system tolerance. Maybe Chris or someone at VX knows a workaround??

A better option, I think, would be to scale the data before importing. It sounds like a .001 scale factor might work, i.e., make 1mm in VX equal to 1 meter in the data. Can you scale the point data before importing into VX? Maybe in a spreadsheet, multiply all the x,y,z values by .001.

VX is very good at creating the surfaces, but it will take some work to decide where one surface should stop and the next begin. Also, if you receive multiple data sets and some sets are not correctly registered to the others, you will have major problems aligning them manually. Other programs have automated point cloud registration functionality (Raindrop Geomagic is a good one).

Steve

Rank: 1

Kevin

Newbie

posts: 0

Registered: 2004-4-26

Message 3 of 13

15-09-2005 01:31 . pm   |   View his/her posts only
Never done anything like this in VX but have done this in VectorWorks Architect using the site model feature which creates a surface from inputted survey points. The surface you get is polygonal. This works very well. I'm not saying VX can't do it but perhaps this is an area where more application specific apps and application specific tools would be more useful?

Rank: 1

ChrisWard2k2

Newbie

posts: 2

Registered: 2011-11-22

Message 4 of 13

21-09-2005 04:49 . pm   |   View his/her posts only
Hi OC.

I think VX will handle the job well. Firstly, you will want to avoid scaling data (since then all of your design will need to be scaled, all the verification scaled, etc - too messy). What you can do is change your general working units to metres (at least, for your "sea bed assembly"). I imagine that the sea bed data consists of a single file of XYZ values. There is an upper limit to the number of points that can be imported (the limit is in the millions). I have no idea how many points you have to deal with, but if the file exceeds the limit, it can be divided into several files as required (I can supply a util to do this). After that, the overall sea bed shape is probably described fairly loosely and can be represented by a point cloud surface, but you will need to pay attention to specific shapes here and there - these will need to be isolated and dealt with separately. Have you got any benchmark data I can see?

Rank: 1

Steve

Newbie

posts: 0

Registered: 2004-3-1

Message 5 of 13

22-09-2005 12:21 . pm   |   View his/her posts only
Chris,

The reason I suggested scaling is because I cannot, for example, view the extents of a line drawn from 0,0,0 to 1500,1500,1500, even if my units are in meters. What am I doing wrong? Do I have to change the system tolerance as well? Won't OCP have the same viewing problem?

Steve

Rank: 1

ChrisWard2k2

Newbie

posts: 2

Registered: 2011-11-22

Message 6 of 13

23-09-2005 08:37 . am   |   View his/her posts only
Hi Steve

Well, it isn't you. I can input and verify a line from 0m,0m,0m to 1500m,1500m,1500m. However, "zoom all" will not fit the line neatly on screen as the view extent input defaults to a maximum 1000m. I'm not sure why this is, since in floating point calculations we are concerned about the number of decimals, and when you are measuring in metres the position of the decimal point is going to be favourable. If the units are mm it seems that there is no extents problem for "high numbers". So, if necessary, we can work "unit less" - use the m values as mm values. I will ask the developers if this really is a finite limitation.

Rank: 1

OldForumPost

Newbie

posts: 0

Registered: 2012-1-14

Message 7 of 13

23-09-2005 12:39 . pm   |   View his/her posts only
Hallo guys

The csv file is attached. The zoom is problematic, as can be seen from the file values.

Adding a surface thru these points was interesting; some worked others not. I would like to be able to the complete surface on the screen though. It will make the 'sale' much, much easier for the rest of the tasks such as bollards, piles, decking etc.

Thanks for the input thus far.

ocp

Rank: 1

Steve

Newbie

posts: 0

Registered: 2004-3-1

Message 8 of 13

23-09-2005 06:35 . pm   |   View his/her posts only
Hi OCP,

What is the format of this data file? I can't seem to find the X,Y,Z coordinates.

Steve

Rank: 1

ChrisWard2k2

Newbie

posts: 2

Registered: 2011-11-22

Message 9 of 13

23-09-2005 07:56 . pm   |   View his/her posts only
Hi OC

There are more than three values per line e.g:

First Line: 676872,8392779,-0.83,,0,7715907,-0.83,,7715907
Last line: 679707,8394328,-28.28,,2835,7717456,-28.28

....and the pattern is inconsistent. So, need to know from you what the value of each column represents.

Rank: 1

OldForumPost

Newbie

posts: 0

Registered: 2012-1-14

Message 10 of 13

24-09-2005 04:34 . pm   |   View his/her posts only
My apology! The wrong file. That was the one i use to 'mess' with and export from.

The correct one is attached.

ocp

Rank: 1

Steve

Newbie

posts: 0

Registered: 2004-3-1

Message 11 of 13

26-09-2005 12:46 . pm   |   View his/her posts only
I had to do a simple transformation of the data in excel to work around the view extents limitation. I subtracted 676000 from all X values and 8392000 from all Y values. Then imported the csv files (with my units set to mm) and created an untrimmed surf through the entire point cloud. The surface is fairly flat.

If you try to undo the transform in VX by moving 676000 in X and 8392000 in Y, you hit the same view limitations, so this is an issue requiring some kind of work around. Chris???

Steve

Rank: 1

ChrisWard2k2

Newbie

posts: 2

Registered: 2011-11-22

Message 12 of 13

26-09-2005 07:22 . pm   |   View his/her posts only
Hi chaps

Yes, it is indeed necessary to get the data closer to the origin. The data does look a bit suspicious too.........certainly not that exciting! I am waiting for a response from Florida because I feel we should be able to handle this file with an increased view extent.

Rank: 1

OldForumPost

Newbie

posts: 0

Registered: 2012-1-14

Message 13 of 13

28-09-2005 03:31 . pm   |   View his/her posts only
Hi Chris

I raised a PCR for the view extents.

The data is good. It is for a seabed. We will be adding the rock formation data soon thus enabling the customer to check the pilon length requirements at specified points. This is proving to be a very cost effective solution (once the zoom extents is resloved) compared to soem other products.

Thanks for the input from everyone. Much appreciated by this Capetonian.

See also
X