Gross Error Detection and Convergence Analysis in Photogrammetric Networks

Robust methods of parameter estimation are often employed in multivariate applications where gross errors aaect the data, however robust methods commonly lack pre-and post-analysis measures enjoyed by least squares estimation. In terms of pre-analysis, we describe a mathematically rigorous method for determining redundancy numbers based on L1-norm minimization and for post-analysis we outline an L1-norm-based method for detecting gross errors in photogrammetric observations. Additionally, we describe a graphical method for interpreting the convergence robustness of nonlinear parameter estimators and apply this method to single photo resection.