Outlier detection in surveying networks
MetadataShow full item record
The paper refers to the robust estimation methods, which allows to eliminate outliers in surveying networks. Network adjustment is performed by the method of least squares. A key problem is the correct selection of weights, resulting from the different standard deviations of observations. In the case of gross errors their impact on the results of the alignment can be minimized by reducing the weight of outstanding observations. The second solution is the elimination of such observations as they were detected and re-alignment this network. In addition to the presentation of the well-known features, damping solution, iterative solution was presented based author idea. The calculation is illustrated on the one-dimensional random variable. Also presented the final results of the flat network adjustment by the proposed algorithm to eliminate outliers.