Extracting toolpath profile from 3D scan of wheel

[marekm]

I'm wondering if anyone here can suggest suitable math steps to extract a 'surface to cut' profile radial line from the test data in the Dropbox folder?

Attached is the vi for viewing the scanned data.

The extracted profile needs to be a radial line that runs across the highest points (the spokes) from the middle of the wheel to the outside edge (standard lathe machine tool trajectory).

There needs to be some level of outlier correction on the data as there is a few erroneous points here and there (obvious in attached 3D viewer).

Scan data and corresponding pictures for various wheels can be viewed here:

https://www.dropbox.com/sh/r9cdmyy5s6iyi6l/0Pomg0dmyp

PS. In the data, +ve Z is downward.

Regards

Marek

[marekm]

Update to avoid out of memory error.

[alexderjuengere]

---

@marekm wrote:

I'm wondering if anyone here can suggest suitable math steps to extract a 'surface to cut' profile radial line from the test data in the Dropbox folder?

Attached is the vi for viewing the scanned data.

The extracted profile needs to be a radial line that runs across the highest points (the spokes) from the middle of the wheel to the outside edge (standard lathe machine tool trajectory).

There needs to be some level of outlier correction on the data as there is a few erroneous points here and there (obvious in attached 3D viewer).

Scan data and corresponding pictures for various wheels can be viewed here:

https://www.dropbox.com/sh/r9cdmyy5s6iyi6l/0Pomg0dmyp

PS. In the data, +ve Z is downward.

Regards

Marek

---

Hi Marek,

first a general hint, as your data is a binary file you should zip your files.

Wheel1.bin is about 100 MB large, but if you zip it, you'll get a compact 13 MB file.

The extracted profile needs to be a radial line that runs across the highest points (the spokes) from the middle of the wheel to the outside edge (standard lathe machine tool trajectory).

Do you expect something like this:

1# Define the center of your object A

2# then define a point B on the objects edge

3# Extract tall xyz points which lay on the line AB ?

Regards,
Alex

[marekm]

Hi Alex, thanks for your reply.

Compressed as recommended!

"Do you expect something like this:

1# Define the center of your object A

2# then define a point B on the objects edge

3# Extract tall xyz points which lay on the line AB ?"

Yes and no. This is a machining process that just skims the top face of the wheel so we are only interested in the highest of points throughout each radial length (I mean we aren't interested in any data down the sides of the spokes or in between the spokes).

The problem is, there are a scattering of spurious measurements that are usually not attached to the main mass (I'm thinking 3D blob processing) to remove before that (radial ring max hold) calculated line can be reliably extracted.

Any ideas?!

[alexderjuengere]

hi marek,

Here's what I had in mind:

1# Extract all the red line's points, use XY Graph to display extracted line

2# Leave those points out which have no hight information (z<0), use XY Graph to display extracted line

I will be away for 3 days, but I'm looking forward to discuss with you

Regards,
Alex

[alexderjuengere]

as requested, here's the code attached