Results and Discussion
This chapter assesses the performance of the Black Spot modules. First, the unknowns that are present in the system are identified in Section 8.1. Following this, in Section 8.2, results are given for the modules tested in isolation while stationary. Results gained by testing the pose algorithm using synthetic data are shown in Section 8.3. Due to the performance of the pose estimation algorithm, results from the entire systems are not presented as it was decided that these could be misleading.
Uncertainties in parts of the system affect the overall performance, therefore it is important to know where these lie. There are two types of unknown; system constants and time varying unknowns. System constants do not change (or change little) during operation.
Examples of constant unknowns are the intrinsic camera parameters, i.e., lens focal length, lens distortion, and image sensor pixel size. While it is possible that these parameters may change slightly, for example, due to change in temperature, it is assumed that this change is small and that they are essentially constant. Careful calibration can remove the effects of this type of uncertainty.
There are two types of time varying unknowns; those that can be modeled and those that cannot. n example of a time varying unknown that could be modeled is the variation in measured intensities of the markers such as a 50 Hz sinusoid from the mains power which
theoretically compensate for this but, drivers.
in this case,
it would be better to improve the LED
Some time variant unknowns cannot easily be modeled. Examples are the non-cyclic noise