Basically we would like to know the demand and interest on the provision of data sets for a certification on software products in close-range photogrammetry.
We would like to know your opinion concerning the considerations of software certification in close-range photogrammetry.
The questionnarie can directly be filled out in the MS-Word document. Beyond a pdf-document is provided for download.
As a result of a questionnaire campaign at the Oldenburger 3D-Tage and a couple of conversations with software producers and users the
following discussions arised:
Is it necessary to provide 2D and 3D data sets for the verification of software? This question was answered in the affirmative.
System configurations are built with respect to the application. Reducing the data material may lead to high differences
and uncertainties. Blunders and a minimum of ray intersections for object points are treated within the bundle adjustment software
in different ways and may significantly change the result. Which scenarios are significant and require verification? The feedback of
the questionnaire campaign does not give a uniform statement. However, different configurations and considerations of blunders are of
the majority noted as necessary.
The sets of parameter of a bundle adjustment require knowingly handling. Control information such as control points and
system scales have to be considered. This kind of information depends on the application. Routines for blunder elimination, with
automatic or manual handling as well as reactivation algorithms for pretended blunders, are implemented in many bundle adjustment
software packages. These intervention possibilities for the user may lead to great differences within the results. May and must
have all these scenarios to be verified? As a result of the discussions: high attention has to be paid on system scales and algorithms
for blunder elimination.
The functional model for the interior orientation within the bundle adjustment may be different. According to the used camera
and lens the parameters of interior orientation affect the analysis and accuracy of the results. To provide data sets for the verification
of software the approaches have to be defined in order to guarantee an appropriate data generation. Current considerations do not consider
affinity and shering. This needs to be discussed furthermore. If necessary different error-free data sets have to be generated and provided.
The discussion on which data will be provided first resulted in the provision of error-free data sets to verify the fundamental
operation of the software.
Beyond that it is possible to provide data sets generated under consideration of uncertainties. Based on the return of adjusted
object points length measurement errors can be determined. For this analysis method it has to be considered if all approximate values or
merely observations and control information were given to the adjustment. Already the calculation of approximate values may cause an
incorrect adjustment and a deformation of object space. The discussions did not lead to a uniform statement.
Impressum der Jade Hochschule – Datenschutzerklärung der Jade Hochschule
Legal Notice of Jade University – Provacy Notice