Accessing and Exporting Marker Error (pix) values

[AndyMatthews]

Hi Guys,

Occasionally Marker detection returns a false positive and misidentifies a coded target. We are reviewing the type an size of the markers to address this issue, but for now we would like to find a way to flag this issue as part of our batch process.

In the Reference tab and within the Markers table we can see the target has an Error (pix) value.

Where there is a false positive, this value is significantly different to the other values.

Using a Python API, how can we read and export the stats on each detected marker. From this we can raise the flags.

Further to this, is there a way to quantify the degree of error for each marker on an image basis, so we could perhaps cull the false positive from the data set all together?

Many thanks

Andy M

[Alexey Pasumansky]

Hello Andy,

There's .pinned method for PhotoScan.Marker.Projection() class variables that returns the status of the marker projection (pinned = green flag, not pinned = blue flag). Gray flag location is not stored in the project and should be calculated according to the relative position of marker in 3D and camera. So actually, basing on the 3D position of the marker (that is computed from the existing projections) you can locate the projection position on 2D image. Error in pixels for the marker projection can be estimated as a difference between this projected location and the position of the flag on the photo.

Let me know if you need further assistance, for example, for the marker projection estimation.