Elsevier Science Home
Computer Physics Communications Program Library
Full text online from Science Direct
Programs in Physics & Physical Chemistry
CPC Home

[Licence| Download | New Version Template] aadh_v1_0.gz(72 Kbytes)
Manuscript Title: CENTER: a software package for center estimation.
Authors: S.B. Hooker, J.W. Brown
Program title: CENTER
Catalogue identifier: AADH_v1_0
Distribution format: gz
Journal reference: Comput. Phys. Commun. 38(1985)421
Programming language: Fortran.
Computer: VAX-11/780.
Operating system: VMS.
RAM: 15K words
Word size: 32
Peripherals: disc.
Keywords: General purpose, Geophysics, Center, Vortex, Harmonic, Fit, Optimization, Switch processing.
Classification: 4.9, 13.

Subprograms used:
Cat Id Title Reference
AADI_v1_0 FIXSRC CPC 38(1985)435

Nature of problem:
An algorithm for estimating the center of a discretely sampled two- dimensional contour in space is presented. In particular, the effects of data sampling problems (i.e., partial sampling, uneven data distribution, feature translation, etc.) on the estimation process are considered. Optimization procedures for preventing degradation in the center estimate as a result of such problems are included.

Solution method:
The center of the contour is estimated by the method of intersecting adjacent perpendicular bisectors. The center estimate is the mean of wanted intersections between perpendicular bisectors constructed between adjacent data point pairs. Wanted intersections are those intersections that are not rejected by the optimization procedures. The methods employed by the optimization routines also allow for estimation of the shape, size, and orientation of the feature.

The primary restriction is not one of complexity as much as one of feature delineation, that is, the sampling employed must adequately resolve the feature if poor center estimation is to be avoided. This is particularly true with features that are quite elliptical. Another problem to avoid is uneven data distribution, the sampling should be relatively evenly spaced. The optimization procedures used tend to minimize the effects of sampling biases, but it is unlikely the procedures will be successful if such problems are very severe.

Unusual features:

Running time:
Compilation: 66 s., execution: 2 s.