Influence of line scale contamination and defects upon calibration uncertainty in dynamic mode

Abstract

This paper presenting new calibration procedure method, which was introduced on precise linear long scale comparator with video detection system. In most cases nonmechanical uncertainty budget consist from line edge detec-tion uncertainty. Line centre is detected though processing signal of digital camera sensor. Our algorithm for line edge approximation and centre detection is used. Edge detectors are more or less susceptible to noise present on raw unpro-cessed image data and different types of line contaminations (dirt spots, digs, scratches). Image noise can be slightly suppressed by various smoothing filters with not significant data information loss. Line contamination can’t be repaired or reconstructed without a priori information about line edge. We suggest method for line incorrectness detection and removing from line edge profile. Images were taken from real glass scale in order to test ability of new method.