Meaning
Metrology performance metrics calculate the average deviation of measured coordinate points from a best-fit geometric model. The residual rms error provides a statistical measure of the noise and inaccuracies remaining after calibration corrections have been applied. This value is derived from the square root of the mean of the squared differences between the observed and predicted values.
It quantifies the quality of a camera calibration or surface scan.
Calibration Quality
A low residual rms error indicates that the mathematical model closely matches the actual physical geometry of the system. In photogrammetry, this metric reflects the alignment precision of the multi-camera network. High errors point to problems with lens distortion correction or target identification.
Mathematical Calculation
Summing the squared distances of all data points from their predicted positions and dividing by the total count yields the mean squared deviation. Taking the square root of this value produces the final metric. This calculation heavily weights large individual discrepancies, making it a highly reliable tool for spotting outliers, sensor noise and systematic distortion.
It is the primary value used to accept or reject a high-precision calibration run.
Geometric Limit
Very low error values can sometimes hide systematic distortions if the data points are poorly distributed. A localized deformation might not register if the rest of the scan has high point density. Analysis of the residual vector map remains necessary to confirm the absence of localized distortions.