Nonlinear Measurement Data Processing Based on Improved Damping Least Squares
Datum are the key of “Digital Earth”.In measurement, dealing with nonlinear models of observation datum, we may take their approximate values at observation values by Taylor series expansion, say, taking first-order item as a linear function of classical adjustment. But requirements of observation data, processing and accuracy assessment are higher and higher with today's fast-growing of high-tech mapping and surveying. So study on nonlinear least squares adjustment has been paid more and more attention. Damping least squares, as a modified algorithm of Gauss-Newton’s algorithm, is necessary to add a damping factor to improve the nature of a coefficient matrix. But it is difficult to choose a suitable damping factor, and needs to solve a group of linear equations repeatedly. In this paper, an improved damping least square was utilized for the non-linear processing of measurement datum in order to reduce a lot of computational workload.
Paul P. Lin and Chunliang Zhang
G. L. Li "Nonlinear Measurement Data Processing Based on Improved Damping Least Squares", Applied Mechanics and Materials, Vols. 105-107, pp. 2034-2038, 2012