Is it possible to use a file with non-grid pattern for autolevel feature?

[muhammadelmogy]

I want to know how does the grid points generated in the Autolevel feature and is it possible that those points are not in a grid pattern?

I am working on a tool to do auto level for PCB and have the ability to avoid drill points so the generated grid points will not exactly in the a grid pattern. something like this:

As you can see some points will be shifted from its original place to avoid probing a drill point and other points may be omitted totally from the grid like the points around the drill points at the corners, you can click on the image and zoom in for more details

So my plan is to use those generated points and create the height map and save the result in a file to use it in ioSender!
Is it possible to use a file with non-grid pattern for Autolevel feature?

[terjeio]

You have to modify the code for that - AFAIKT the current one relies on the fact that the points are at an equal spacing.

[muhammadelmogy]

I was really hoping I could do this without touching ioSender since the idea is to generate the height map file before using ioSender and just let ioSender use it.

I have already finished writing the code to do all of that starting from parsing Excellon file, generate a grid, make the necessary modifications
to the generated grid and then use it to probe a PCB surface and finally output a height map file.

I was thinking of making several versions of the output file so it would be used in other software and that would make it more versatile.

[muhammadelmogy]

If I understand correcly then you have the probed distances already?

Yes, the provided height map file contains all necessary data. It features a modified grid pattern where each point has a probed distance associated with it. The first point is set to zero, and subsequent points are referenced to it by subtracting the reference distance from their z-distance (z_distance – reference_distance).
The process begins with generating a regular grid based on user input values for the x and y coordinates, including the minimum and maximum values, and the number of points in each direction (x_min, x_max, y_min, y_max, x_points, y_points). If an Excellon file is provided, it is then parsed to extract drill point data, including the x and y coordinates, as well as the drill size (x, y, drill_size).
The grid points are then compared to the drill points, resulting in one of three possible states for each grid point: no change, shifted within a specified range, or omitted. The resulting modified grid points are used to probe the surface of the PCB, generating a height map file.

[terjeio]

The resulting modified grid points are used to probe the surface of the PCB, generating a height map file.

So the probe values are not the ones to be used? Only the XY coordinates?

[muhammadelmogy]

The resulting modified grid points are used to probe the surface of the PCB, generating a height map file.

So the probe values are not the ones to be used? Only the XY coordinates?

The final generated file from my side have x, y and probed value
This final file can be used directly without the need to make a new height map anywhere else (or that's what I am hoping for) I was hopping to load it in the height map and use it directly.

[terjeio]

This final file can be used directly without the need to make a new height map anywhere else (or that's what I am hoping for) I was hopping to load it in the height map and use it directly

Ok, then I believe the interpolation code linked above has to be changed. You can do that?

[muhammadelmogy]

This final file can be used directly without the need to make a new height map anywhere else (or that's what I am hoping for) I was hopping to load it in the height map and use it directly

Ok, then I believe the interpolation code linked above has to be changed. You can do that?

After reading about different kinds of interpolation techniques, I found that I do not know anything about them :D 🤣
but one thing I found those 3 methods

Interpolation methods designed for irregularly spaced data. Common techniques include:
Triangulated Irregular Network (TIN): Creates a mesh of triangles from your actual data points and interpolates within the triangle containing your target point (e.g., using barycentric coordinates).
Inverse Distance Weighting (IDW): Estimates a value at a point by taking a weighted average of known values, where the weight decreases with distance from the known point.
Kriging: A more advanced geostatistical method that considers the spatial autocorrelation of the data.

To answer your question: short answer no I can't do it. long answer yes if I have enough time to understand and learn the answers for how/when/why to use those techniques.