Tie Point Accuracy setting in Ground Control Settings

[andyroo]

I was wondering if anyone collecting aerial imagery has experimented with tie point accuracy settings. My default is 4 pixels, and that seems a little large, but I don't know anything about that variable.

Andy

[Alexey Pasumansky]

Hello Andy,

In fact tie point accuracy parameter effects on the weight of tie points during optimization process. Actually, changing this parameter is equal to changing all the others parameters in the opposite direction, so we do not recommend to change it except some extraordinary cases - marker/camera/scale-bar weights should be used.

[macsurveyr]

I really think this needs more discussion.

As with all least squares analysis, the weight - the error estimates - for any and all measurements - the sources of error - need to be correct in order for the results, and just as importantly, the statistics quantifying the results to be correct.

I do not believe that a 4 pixel error is a valid estimate for the tie points image matching accuracy.

We can have a long discussion about what it should, or rather could be, depending on the matching algorithms used, but it is quite common to expect sub-pixel matching - sub-pixel error - even as low as 0.1 to 0.2 of a pixel of error. In PhotoScan Pro with minimal optimization and gradual selection removal of the worst points, a point variance RMSE of less than 0.5 pixel is very easy to obtain. So, a 4 pixel error estimate is very pessimistic. An unreasonably pessimistic error estimate.

It also seems to me that entering a much more realistic estimate - in the sub-pixel range - does not in fact change the statistics as much as it should. It still seems that control, whether it is true ground control or camera position values, are given too much weight not necessarily honoring the entered values.

Tom

[sjharwin]

I am with Tom on the need for clarification of the Ground Control Settings - Measurement accuracy settings.

I am assuming a few things about these settings as I cannot find much info in the WIKI or in the help that defines the meaning of these values. My fist assumption is that the accuracy value is 1 sigma. My second assumption (based on reading the forum) is that I should leave the tie point accuracy set to 4 pixels (which seems large as Tom said). My final assumption is that the Projection accuracy value is an indication of how well a well defined feature can be pinpointed in an image and therefore in the model... which I would have thought should similar to the tie point accuracy. In general a human can identify a well defined feature to around 0.4-0.6 of pixel and a computer can do it to around 0.1 and that is why I assume you provide 0.1 as the default.

If you could provide clarification on these settings and how the default values are derived I would be appreciative. My GCPs have been measured with a precise total station survey and therefore I can confidently say they are good to a sigma 1-1.5 mm. In one of the forum posts you suggest that I should therefore set the Marker accuracy to 0... is this really the best approach?

Thanks
Steve

[ann]

Hello everyone,

I'd like to refresh the topic about tie point accuracy.

As Alexey Pasumansky wrote:
"In fact tie point accuracy parameter effects on the weight of tie points during optimization process. Actually, changing this parameter is equal to changing all the others parameters in the opposite direction, so we do not recommend to change it except some extraordinary cases - marker/camera/scale-bar weights should be used."

Are all accuracy parameters just weights in fact?
How are they taken to calculations? In direct proportion? Inverse?
Alexey, could you elaborate on that?

Anna