Abstract:
|
EIT acquisition systems, when regarded as measuring instruments, can be
characterized by their inherent characteristics: reproducibility, repeatability and discrimination.
The ultimate limit to all three is noise, whether long-term or short-term noise. In addition, for
some applications accuracy can also play a role, although accuracy is not an inherent
characteristic and requires a known test object to be defined. Because EIT systems are
multichannel instruments and the information of all channels is being combined to produce an
image, a procedure to standardize the performance assessment must be set up in order to allow
for meaningful comparison of different systems or different settings in the same system. We
propose the use of three parameters defined long ago but scarcely used among EIT systems
designers: SER (Systematic Error Ratio), NER (Noise Error Ratio) and RER (Reciprocity Error
Ratio). We applied these definitions to our latest EIT instruments, TIE5-sys, in a wide
frequency range and with different settings in configurable acquisition parameters and provide
the interpretation of the results obtained in terms of reproducibility, repeatability and
discrimination. We also comment on the tradeoffs that show up when using the usual definition
for these figures |