Re : Cursor crosses curve in 2 points: find 2nd point

[Pavel_47]

Hello,

Here is emulation of a part of my mesuerement setup.

Here are keypoints:

• measurement curve (emulated by sinc function)
• 2 free cursors TOP and BOTTOM
• 1 locked cursor "MIDDLE": its Y is calculated as mean value of TOP and BOTTOM
• 1 crossing point can be derived from Active Cursor property (PosX and PosY)

Does exist a simple way to find 2nd crossing point.

Thanks in advance

[Pavel_47]

Here is a solution which is a little bit tricky.

It works as follows:

1. On cursor release (TOP or BOTTOM) the mean value of Y is calculated
2. Cursor MIDDLE (locked to the curve) is set
3. Coordinate X of the cursor MIDDLE is derived
4. Index of the coordinate X is found in the original array
5. Original arrays X and Y is cut next to X coordinate index
6. New curve is built on supplementary graph "XY graph 2"
7. Coordinate X of the cursor on supplementary graph is found
8. Coordinate (X1, Y) and (X2, Y) are applied to Annotation List collection

The probelem here is when I hide "XY graph 2", the method isn't stable - sometimes original curve is cut correctly, sometimes - NOT.

Any comments ?

[Pavel_47]

Here is approach (still a bit complicated) that works:

[alexderjuengere]

this here would get more than two crossings

[alexderjuengere · Accepted solution]

Ok, sorry for hijacking this thread, case "all" was quite easy:

0# all:

getting the first x-position wasn't too bad

first:

getting the last x-position of a crossing was slighty harder but feasible without queues:

last:

....but getting the middle was quite tough, is there an alternative besides using queues ?

median:

Regards,
Alex

[Pavel_47]

Hello Alex,

Your approach is quite ingenious. Congratulations !

Especially I liked "median" case, actually it's what I need.

Thanks

[alexderjuengere]

---

@Pavel_47 wrote:

Hello Alex,

Your approach is quite ingenious. Congratulations !

Especially I liked "median" case, actually it's what I need.

Thanks

---

you are welcome 🙂

This version of the median case mighty be a bit more comprehensible:

Regards,
Alex