How is the total marker error (pix) being calculated

[colin.spexigeo]

Hi there,

I'm interested in how the "Total" value of marker errors is being calculated. I had assumed that it was an average (mean) of all camera projection errors, but this does not seem to be the case.

For example, Agisoft reports the following:

Errors
-dYUxAmaeU	0.143
-YDOHdBJSz	0.458
-YOo6Otjv5	0.629
CYzKGPFuwI	0.437
f6n4TJCcSD	0.054
FqIO_ToeyF	0.397
Fy4qsbY3ft	0.661
ladYl1TVsF	0.497
nziQ02gyo5	0.322
pfZjrVEIQA	0.518
rFtTrwWuiN	1.332
tQGuj6cN-y	0.585
VcznBnOJaW	0.358
XDxBg3CLSh	0.135
Total	0.554 pix

However, the *average* values of all these errors is actually 0.466.

Does anybody know how the "total" values is being calculated?

Thanks in advance,
Colin

[Paulo]

The total error is the RMS of  marker errors

Marker           Error
dYUxAmaeU   0.143
-YDOHdBJSz   0.458
-YOo6Otjv5   0.629
CYzKGPFuwI   0.437
f6n4TJCcSD   0.054
FqIO_ToeyF   0.397
Fy4qsbY3ft   0.661
ladYl1TVsF   0.497
nziQ02gyo5   0.322
pfZjrVEIQA   0.518
rFtTrwWuiN   1.332
tQGuj6cN-y   0.585
VcznBnOJaW   0.358
XDxBg3CLSh   0.135
RMS           0.554  SQRT(sum.squares(B2:B15)/14)

[colin.spexigeo]

Thank you Paulo!

P.S. Off-hand, you wouldn't happen to know an easy way to calculate the "Error (m)" value, would you? I imagine you'd have to project the marker points/error offsets through the cameras (maybe projected onto a plane that's facing the sensor) and run an RMS on the cartesian-space error values. I'm trying to calculate this information and surface it to the user outside of the program.

[Paulo]

Assuming the markers have reference locations, then individual error for each marker is calculated as:

Code: [Select]

for marker in chunk.markers:
      source = marker.reference.location
      estim = chunk.crs.project(chunk.transform.matrix.mulp(marker.position))
      error = estim - source
      total = error.norm()
      print(marker.label, error.x, error.y, error.z, total)
and total error again is calculated as RMS of each marker error...Example:

Code: [Select]

Marker Error.X(m)  Error.Y(m) Error.Z(m)   Error(m) Error(pix)
PA03 0.077        0.047 0.017        0.092 0.120
PA08 0.041       -0.068 -0.020        0.082 0.302
PA12 -0.052        0.031 -0.005        0.061 0.254
PA18 -0.053        0.019 -0.005        0.056 0.086
PA01 -0.016       -0.041 -0.004        0.044 0.228
PA19 0.006       -0.022 0.017        0.029 0.094
PA16 -0.017       -0.006 -0.005        0.019 0.135
PA11 0.007        0.014 0.010        0.019 0.117
              RMS      0.056    0.184