Non-destructive Detection System for Micro-particle Contaminant on Super-smooth Optical Surface

摘要

Regarding the demand of a high cleanness optical surface for an experimental prototype, a method of nondestructive detection of micro-particle contaminants is proposed. Aimed at the detection of cleanness, an improved, self-designed scanner image sampling of the particle contaminations on the optical surface was used. The hardware elements implemented with the algorithm of image processing to detect particle pollutant, based on the Canny's operator, in order to identify the edge of particle. The construction the local closed interval was done by calculating a convex hull with Graham algorithm, thus obtaining the image of particle contamination via the threshold segmentation. Identification of the particle contamination based on RVM (Relative Vector Machine) is demonstrated. The results from the relevant experiments in the super-cleanness circumstance confirmed the usefulness of this system.