Discussion on 'Jump regression, image processing, and quality control'

摘要

I very much welcome this opportunity to make some comments relating to the article by Professor Qiu. I am a mathematician at the National Physical Laboratory (NPL), the UK's National Metrology Institute (NMI), the institute responsible for ensuring confidence in measurement throughout the UK. Confidence in measurement comes by (i) agreeing a standard set of units, the metre, kilogram, etc., (ii) developing primary standards that realise these units, and (iii) putting in place calibration methodologies that allow test artifacts to be calibrated against standard artefacts, producing a chain of calibrations from the primary standards down to working standards used to check measurements made by industry and commerce. An important element in ensuring traceability to standard units is the evaluation of measurement uncertainty: a measured value without an accompanying statement of uncertainty is not regarded as a valid measurement within the metrology community.