At present, new detection techniques such as image-based detection techniques, have become an indispensable feature in modern industrial production. During mass industrial production operations, image-based detection ensures the consistency of product detection and helps implement data quality monitoring and process control, thereby increasing detection security, reliability, efficiency and precision, and reducing production costs. According to the different characteristics detected image-based detection applications are categorized as dimension measurement, surface quality detection, structural quality detection and system operating status monitoring [1]. Among these applications, dimension measurement and surface quality detection are the most commonly used.

Dimension measurement mainly involves characteristics of a target such as appearance, shape and position. It is also used in other fields, such as detection of discontinuous arc roundness in the field of machining [2], assemblage clearance in the field of automotive industry [3], and excursion and deflection of chips in the field of electronics, such as in printed circuit board (PCB) production [4]. Surface detection mainly involves detection of defects that impact product surface quality, such as fovea, scratches, cracks, air bladders, holes, wear, roughness, texture, and burrs, such as in steel plate surface defect detection [5], surface roughness measurement [6], tunnel wall surface defect detection [7], welding seam defect detection [8], fabric surface defect detection [9], and wood defect detection [10].

In this paper an image-based detection technique used for pivot bearing dimension measurement and surface defect detection is described. Illumination system design has a direct relationship with final imaging quality, and it is one of Carfilzomib the keys for the success of any vision detection system. Improper AV-951 illumination may give rise to many problems; for example, overexposure may hide true defects, shadows may cause edge false drops, and non-uniform illumination may cause image segmentation difficulties. As a result, illumination quality can directly impact image analysis results [11].