A Fine Technique for Automatic Inspection of Surface Mount Components

Abstract

In this paper, a machine vision technology is used to inspect surface mount components (SMC). It can receive the purpose of improving manufacturing losses caused by misjudgments of manual inspection. The detection includes missing parts, wrong parts, skew, reverse polarity, missing pin, bridging, broken pin, etc. The surface mount components to be tested in this paper include QFP (Quad Flat Package) and SOP (Small Outline Package). The proposed method consists of four parts: (1) image capture: capture the printed circuit board (PCB) image in XY-Table through camera; (2) detection: use correlation coefficient to detect missing and wrong parts; (3) positioning: judge from the RGB histogram of the shifting element; (4) segmentation: detect disconnection and bridging defects according to the segmentation result of the cumulative projection.