Non linear curve fitting

[Panjimar]

Hi ,

I have a set of data points , using these points I need fit non linear curve. I don't know which VI I have to use for fitting non linear curve. 

What are the parameters use for the VI? and How to decide the parameters for VI?

If  anybody knows please let me know.

[IanSh]

Function palette>mathematics>fitting.

Take your pick. The relevant help pages and examples will probably give you the info you need on parameters etc.

Also 

Help>Find examples...Search>"fitting"

I found 19 examples there, hopefully you can find what you need.

[Gregory]

Do you have any functional form that you want to fit to? Or do you just want an Nth order polynomial?

[altenbach]

Except for a straight line, every single model is nonlinear, so you should probably be a bit more specific. How does the data look like? Are you trying to fit to a simple model (polynomial, exponential, etc.) or something more complicated? Do you need to get physically relevant parameters or are you just trying to show the data behavior with a smoother curve?

As a start, have a look at my NI-Week 2017 presentation about data fitting. Then see if any of the stuff in my group about nonlinear fitting will help.

If you are still confused, please come back and ask a more specific question. Don't forget to add your incomplete VI and some typical data. Also give a reference describing the mathematical model you are trying to fit.

[Panjimar]

Hi, 

I am measuring plastic object size using camera which is mounted at certain height and an angle. The camera capture the object images continuously. The object entering the field of view of the camera at certain height and fall down to the ground and again jumps and again falling continuously.

I am finding the object size at different height using image processing. But due to defocus and some other noise I am not able to detect correct object size at some places in the camera field of view. Using available data I need to find correct or approximation size of the object at different heights. For that I am expecting solution from non linear curve fit. But I am not yet find any solution. Please help me to find solution

[altenbach]

Can you show us an example picture and what you are trying to get out of it?

[Panjimar]

Can I share the data which include object location(x,y) and Object size? Right I don't have images

[altenbach]

Unless the object is a sphere, the size and shape will critically depend on the orientation. How do you tell object pixels from background pixels (color? etc.).

You really need to give a more detailed problem description.

So I guess you want to do a 2D fit to a projected filled circle with (1) xcenter, (2) ycenter, (3) diameter as fitting parameters. Is this correct?

---

@Panjimar wrote:

Can I share the data which include object location(x,y) and Object size? Right I don't have images

---

If you already have the results, we no longer need to fit. Do you want to fit the trajectory to e.g. a parabola? Determine the gravitational constant?

We clearly need more information!

[Panjimar]

Hi,

The object is in white color and the back ground is black color and the images captured using gray camera. I the object look like sphere. I need to find the object size at different heights. I shared the VI which include object location(X,Y) and size of object. Please take a look.

The aim of my project is to find how many times the object touch surface. I know the object size at surface. But due to noise the object size at certain height also becomes at surface object size.

[altenbach]

---

@Panjimar wrote:

I need to find the object size at different heights.

---

Which coordinate is "height"? You already have the size in this data. Is this a calibration run? Do you want to get a polynomial describing e.g.  size as a function of y that you later can apply to other measurements? What is the purpose of the x data?