Terrain / Ore Scanners scan direction skews when rolled along its scan axis

[obtuse turtle]

I've been trying to write an algorithm to compensate for vehicle roll and movement while scanning (allowing for multi-pass scans in motion) and I've noticed that the scanners don't always scan directly along their forward axis. While testing, I put a scanner on a (nearly) level static base (with hinges for pitch and roll) and found the following:

• ✅ Pitching up/down with no roll gives consistent distance readings to terrain or ore (although I'd prefer the distance readings to features to increase with pitch. Feels more like a scanner and less "gamey").
• ✅ Rolling left/right with no pitch gives consistent readings as well.
• ❌ Rolling left/right (around the sensors forward axis) while pitched up/down skews the direction of the scan, warping any map generated from the data. I would expect the roll not to affect the forward scan direction. At high pitch/roll angles, this becomes very substantial.

[river pine]

Can you send us a blueprint?

[river pine]

@obtuse turtle

[obtuse turtle]

Here is the "quick and dirty" test platform I'd made. Just place on flat ground and rotate "heading angle" so scanner faces toward the ocean or other notable terrain feature. Right side of screen will show elevation map of scanner readings, with nearest readings at the top. Height scale and range are configured in code (no ui).

[obtuse turtle]

@river pine Let me know if you have trouble duplicating this behavior, or if it is supposed to work this way and I'm misunderstanding something.

[tired yacht]

As a side note of this

It stretches out due to the fact that it isnt wrapping around the planet

[obtuse turtle]

I'm not sure I understand what you mean. Are you referring to long range (planet sized) scans, and how they "unwrap" the spherical coordinates to the flat scan/screen?

[tired yacht]

I mean

Yea

I'm more questioning myself on that

[obtuse turtle]

I've suspected that's how it works, with scans of the other side of the planet being stretched around the outside of a scan. I've not really tested it though.

Also, after further testing on my scanner rig, it appears that as the roll angle of the scanner approaches 90 degrees (on its side), if the scanner is pointed above the horizon the actual scan direction is in the direction of the bottom of the scanner (up to 90 deg off from forward). If the scanner is pointed below the horizon, the actual scan direction is in the direction of the top of the scanner (also 90 deg). Less roll angle = "more forward" scan.

[upbeat shuttle]

Makes sense if the scan only looks in a direction based on a tangential plane to the surface of the planet instead of the vehicle orientation.