On 8/27/2016 2:27 PM, whit3rd wrote:
On Friday, August 26, 2016 at 10:47:41 PM UTC-7, Leon wrote:
whit3rd wrote:
On Friday, August 26, 2016 at 8:53:15 AM UTC-7, SBH wrote:
Anyone see this yet?

https://www.youtube.com/watch?v=_0clzXWSaCw

(3) it registers on tapes that you apply, presumably they have to be
straight and on a perfectly flat surface? Won't happen outdoors in wind, or
on a jobsite with sawhorses under a sheet of plywood.


Videos reveal tat the tape can be crooked, not parallel to other strips,
and the table does not have to be straight. The unit is mapping the dots
and memorizing where they are.

If the surface isn't flat, the map when viewed from one point will not
conform to the map when viewed from another point, against a model
that infers the X - Y coordinates in a plane from those images.

Yes, that is to be expected with most any CNC. And I will correct I
mentioned that the table did not have to be straight, I meant to say
that the tape did not have to be straight.





The vision model uses multiple dots , and it may (or may not) care whether
they are aligned or along a curve; my concern was that a tape pulled taut (in the test)
makes a linear array of those dots, but a gust of wind could disturb the tape
at time of application.

I suspect that the tape has an adhesive back and wind should not affect it.