Abstract
Sets of time domain reflectometry waveforms analyzed with appropriate chemometric methods, instead of typical single waveform analysis, enable one to detect and quantify differences between waveforms, even if they appear at the level of several orders of magnitude below the scale of the measurements. Typical applications of such an approach to samples of healthy and damaged grain showed that differences contributing five orders of magnitude below the scale are still detectable. Under more rigorous experimental conditions, the level of detectable differences can be expected to be even lower.
