Jigsaw Constrain Parameters Not Functioning

[AaronBoyd]

ISIS version(s) affected: 3.5.1 to current (3.9.1) Description
Jigsaw does not respect the constraints placed on points when the point type is set to constrained and a sigma value (point_latitude_sigma, point_longitude_sigma, and point_radius_sigma) is used to constrain the point movement.

How to reproduce

if the user sets the constraint to 0.5, jigsaw might still move the point 50 m

Possible Solution

Additional context
This is a real problem when trying to place images on the surface with known errors, but the solves are not using the limits assigned.

[lwellerastro]

You might consider setting PointType=Fixed if you don't want it to move. That is a pretty tight constraint.

That being said, the constraints set via point_longitude, latitude and radius_sigma are applied to Free points only, not constrained. Constrained points need to have their constraints set manually via qnet by selecting the PointId in the navigator and selecting "Set Apriori/Sigmas" (you can select more than one point at a time and set their sigmas all at once if they will have the same uncertainties). You will only be able to set sigmas here if your points are Constrained (not Free or Fixed), otherwise that section is disabled.

Keep in mind that if you set the constraints too tight (very small values) then your measures are likely to have higher residuals. If you think you know your point lat/lon/radius very well, then just set PointType=Fixed. These also may result in larger residuals. The bundle adjustment will do what it has to to reduce errors overall and if it wants to adjust some but the constrained points are too tight (maybe even inaccurate?) or there are fixed points, residuals overall may go up.

[jessemapel]

@AaronBoyd How were you setting the point constraints? Were you updating them internally in the network or were you passing them through the jigsaw arguments?

The sigma values are just apriori uncertainties and not hard constraints. They are converted into weights in the solution which limits but does not absolutely restrict how far points can be moved.

[AaronBoyd]

We have used both fixed and constrained point types. As a work-around we have used fixed points, but the fact that the constrained points are not using the sigma's that were assigned is the issue being reported here.

The constrained point sigmas were updated in the control network, and also using the command line was attempted as it was not clear that this would not work from the software manual.

Regardless of how the constrained values were entered, the value that the points were moved far exceeded both the cnet sigma values, and the command line sigma values.

The ideal situation with using constrained points and limiting the points to move with different latitude and longitude sigmas is not currently able to be achieved.

[jlaura]

@AaronBoyd After a call with the LROC team, we have hopefully clarified that fixed and constrained points are behaving as mathematically intended. The sigmas are weights in the solution and not fixed ground distance constraints. Have you seen the meeting notes from Megan H.?

[ascbot]

I am a bot that cleans up old issues that do not have activity.

This issue has not received feedback in the last six months. I am going to add the inactive label to this issue. If this is still a pertinent issue, please add a comment or add an emoji to an existing comment.

I will post again in five months with another reminder and will close this issue on it's birthday unless it has some activity.

[AaronBoyd]

Thank you for the clarification