|
When acquiring a new imaging system and operational task performance is a critical factor for success, it is necessary to
specify minimum acceptance requirements that need to be met using a sensor performance model and/or performance
tests. Currently, there exist a variety of models and test from different origin (defense, security, road safety, optometry)
and they all do different predictions. This study reviews a number of frequently used methods and shows the effects that
small changes in procedure or threshold criteria can have on the outcome of a test. For example, a system may meet the
acceptance requirements but not satisfy the needs for the operational task, or the choice of test may determine the rank
order of candidate sensors.
The goal of the paper is to make people aware of the pitfalls associated with the acquisition process, by i) illustrating
potential tricks to have a system accepted that is actually not suited for the operational task, and ii) providing tips to
avoid this unwanted situation.
|
