Two kinds of gross error detection method based on robust estimation
Wen‐Shen Liu · Journal of Liaoning Technical University · 2016
In the case of normal distribution of the measured data, least squares estimation has optimized statistical properties, which is the most widely used estimation method. For the measured data deviates from the normal distribution due to the defect errors, the least squares estimation of error does not have the ability to resist outliers. It is needed to seek estimated theories and methods with the ability of resisting gross errors, identify and eliminate the gross errors of measurement data, so that they do not affect the measurement results. This paper uses two kinds of gross error detection method, the selecting weight iteration method and the method of data detection method, to testing gross errors existing in the observed data. And the standard leveling network data is used to verify the effectiveness of resisting gross errors, which indicts that IGG of iterative method is much easier to realize the detection and localization of gross errors.