Add an option for hitbased construction of features

Jörgen Andersson (vpt se) 8 years ago in Metrology Software / PC-DMIS updated 2 years ago 20 1 duplicate

Add a checkbox to construction dialogs allowing the user to select whether or not to construct the new feature based on hits (<featurename>.HIT[1..<featurename>.NUMHITS]) or the "regular" way (centroids). This would enhance the programming experience and shave off a couple of minutes to the programming time.

Satisfaction mark by Jörgen Andersson (vpt se) 7 years ago
Duplicates 1

If you hadn't suggested this, I would have! Supported.

As it is now, PC-DMIS silently uses hit-based construction for "CYLINDER from CIRCLES" but nowhere else, this is inconsistent.

A possibility to select hits would make it so much more obvious, and would facilitate a lot of constructions which today are more or less cumbersome, for example a plane from the hits of multiple planes.

Some ideas on how this expressiveness could be done are

  • SET1.CENTROID – The centroid of SET1
  • SET1.CENTROIDS – The centroids of the inputs to SET1
  • CIR1 – The centroid of circle CIR1
  • CIR1.CENTROID – The centroid of CIR1
  • CIR1.HITS – All the hits of CIR1
  • CIR1.HIT[1..5] – Hits 1 through 5 of CIR1 (currently supported)

What about those proposition today ?

What is working and what he need to finish ?

Five year later, Can you summurise the situation ?

Many customers have been waiting for ! Thanks for us

I definitely support this as well. I'd recommend (strongly) that it be expanded upon slightly. Counting hit points is an inherently weak approach. In the case of a cylinder measured with 3 levels of hits, I should be able to specify the hits to use "by level". This way, when the number of hits per level is changed, the features constructed from the cylinder's hit points (circles) will update automatically.

So neil.kay, I guess I would add to your list above CYL.ROW[n]

While we're in .HITS-land, I have tried to fetch the surface hits from an autofeature but haven't been very successful in it. There are times when I would like to use them for other things than related to the autofeature, for example when I do measurement of bosses with threads in them, I would like to use the surface hits for the autocircle to evaluate profile (for example) for all the bosses. That would spare me an additional measurement of the surface of the bosses.

Maybe this would better be off in it's own suggestion thread?

I think in this case it's related enough to the initial request to be fine on this thread.

Hi Jorgen, I have been able to edit your script to extract the Sample Hits from Auto Circles. I believe thats what you are describing as Surface Hits. This would be another example to added to "The option to extract individual point data from auto features" Topic.

Thanks Dave, yes - I am talking about the sample hits. :)

I cannot believe this option does not exist within PC-DMIS...or maybe I can. Why is there not a construction option, similar to a line construction from a scan, so that we can easily create a point from any part of a scan? This is a no brainer to me. I would use this all the time. For example, point to point distance or position of thickness or width features. Currently I do not know of any other option that allows me to measure across discrete points similar to a caliper or micrometer would do.

Although I have never used a 3D scanner I think this technology will become more demanded as that technology advances. As I stated before, since I am now using a white light sensor for scanning very often it would seem reasonable to "pluck" out certain data points within that scan without writing multiple lines of code. Whether it be by the click of a mouse or in a pattern type instance to extract that point. I just want to make sure this is not currently possible correct? It seems like it may be with CAD but I also scan without CAD often as well. I just want to make sure that this isn't possible already since I have searched for this tool but have yet to discover it with my current machine.

An example of a scenario for me would be:

I am scanning a ball grid array of solder bumps. I want to extract ever apex of each bump within a certain x,y range but ideally would want to achieve this in 1 scan instead of hundreds of individual ones for a few reasons. 1. Run Time 2. Routine Simplicity 3. Easier re-write/manipulation/extraction of data. I want to extract hundreds of points basically cut into sections along that 1 scan. Although very doable like I said with a lot of code from what I have found, of course a click of the mouse would be prefered. It seems like this could be an entirely new section added onto scanning that could go a long ways if done correctly. Though I do not have any experience with it yet, I think this would work hand in hand with pointcloud? as well as some other PC-Dmis features.

We believe our solution for exposing hit data for construction addresses this request, so will merge the two threads together. This is targeting 2018 R1.

Neil - this topic does not show up as "started" when looking at my profile stats. Yet, it has been marked as "started". Any clues to why?

No. What status does it show? I googled on the UserEcho site, but don't see any reported issues in that area.

It seems to count where I have placed my votes, not how "successful" my suggestions have been, so it seems that it is working as advertised... Too bad though, it would have been nice to see the "mileage" of our suggestions.


The graphical hit selection during construction, dimensioning and alignments will be in 2018 R1, and from the feedback we've had from customers, we believe it should be a very useful tool.  

For sure, there is more we could do here in the future, specifically tools to help find "max" point. This could be achieved utilizing a CAD "Widget" (similar to the QuickSet widget that will ship in 2018 R1 to allow strategy changes), which could offer graphical macros to help select points based on specific scenarios. I think it best to treat that as a different request, so we can resolve this one, which fulfilled the initial request, and let people get their votes back.