Research and implementation of ceramic valve spool surface defect detection system based on region and multilevel optimisation

Abstract

ABSTRACTA sub-regional and multilevel ceramic valve spool (CVS) surface defects algorithm is proposed to improve the adaptation of the existing surface defect detection algorithm and reduce missing rate. According to the different reflectivities in CVS surface, the surface was divided into two regions: polished surface (P) and non-polished surface (Non P), by the characteristics of CVS surface defect, it was divided into two types of defect: G defects and C defects. The P region and Non P region are extracted by searching template image pyramid information, adjusting the target grey-scale image position and posture, regional registration. The defects of the P region and Non P region are extracted by blob detection and crack detection. Through experimental research, the feasibility and applicability of the detection algorithm were verified. The experimental results showed that the detection algorithm was suitable for different types of CVS surface defects, and the missing rate was low, and it can improve the detection efficiency through the reasonable setting of priority.