ISSN:
1573-1626
Source:
Springer Online Journal Archives 1860-2000
Topics:
Architecture, Civil Engineering, Surveying
,
Geosciences
,
Physics
Notes:
Summary A relatively large network is one which covers a large area whereas each observation only involves points within a small local area. For such networks an efficient method is described to establish the lower bounds for the spectral norm of the covariance matrix of the rigorously adjusted co-ordinates. The method utilizes a classification of observations and networks which is thought to be interesting in itself. The weakness of pure angular networks is revealed to stem from the fact that such networks offer little resistance to conformal distortions of the network. A numerical example is given. Asymptotic results for block-type networks are obtained which apply among others to photogrammetric bundle block networks.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1007/BF01614024
Permalink