[ < ] | [ > ] | [ << ] | [ Up ] | [ >> ] | [Top] | [Contents] | [Index] | [ ? ] |
For computation of coefficients in system (1) (ie. during linearization) we need, first of all, an estimate of approximate coordinates of points and approximate values of orientations of observed directions sets.
Approximate values of unknown parameters are usually not known and we
have to compute them from the available observations. For approximate
value of orientation program gama-local
uses median of all estimates
from the given set of directions to the points with known coordinates.
Median is less sensitive to outliers than arithmetic mean which is
normally used for approximate estimate of orientations
During the phase of computation of approximate coordinate of points,
program gama-local
walks through the list of computed points
and for each point gathers all determining elements pointing to
points with known or previously computed coordinates.
Determining elements are
distance between given and computed points
For all combinations of determining elements program gama-local
computes intersections and estimates approximate coordinates as the
median of all available solutions.
If at least one point was resolved while iterating through the list, the whole cycle is repeated.
If no more coordinates can be solved using intersections and points with unknown coordinates are remaining, program tries to compute coordinates of unresolved points in a local coordinates system and obtain their coordinates using similarity transformation. If a transformation succeeds to resolve coordinates at least one computed point and there are still some points without coordinates left, the whole process is repeated. Classes for computation of approximate coordinates have been written by Jiri Vesely.
If program gama-local
fails to compute approximate coordinates
of some of the network points, they are eliminated from the
adjustment and they are listed in the output listing.
With the outlined strategy, program gama-local
is able to estimate
approximate coordinates in most of the cases we normally meet in
surveying profession. Still there are cases in which the solution fails.
One example is an inserted horizontal traverse with sets of observed
direction on both ends but without a connecting observed distance. The
solution of approximate coordinates can fail when there is a number of
gross error for example resulting from confusion of point
identifications but in normal situations, leaving computation of
approximate coordinates on program gama-local
is recommended.
Computation of approximate coordinates of points ************************************************ Number of points with given coordinates: 2 Number of solved points : 2 Number of observations : 4 ----------------------------------------------------- Successfully solved points : 0 Remaining unsolved points : 2 List of unresolved points ************************* 422 424 |
[ < ] | [ > ] | [ << ] | [ Up ] | [ >> ] |
This document was generated by Ales Cepek on June 26, 2017 using texi2html 1.82.