Illumination Optimization for Quasi-Tombstone Detection

Author(s): Mihaly Janoczki, Andras Borbiro, Sandor Nagy, Laszlo Jakab.

Journal Name: Micro and Nanosystems

Volume 2 , Issue 3 , 2010

Become EABM
Become Reviewer

Abstract:

In this paper we describe a novel approach to optimizing the illumination of Automatic Optical Inspection (AOI) appliances for quasi-tombstone detection. At present, optical testers use LEDs (Light Emitting Diode) for lighting. The properties of LED make it a perfect choice for fast, accurate lighting. Complex illumination systems have been developed to reliably detect each kind of failure; but even so these illumination devices are not perfect. At this present time they are not able to provide illumination for each kind of failure using its optimal light structure (that gives the maximum difference between good and bad parts). For AOI systems (and in general for optical sensing), illumination has primary importance. It determines what can be seen on the image when taken from the actual scene. So, the efficiency of failure detection is dependent on the appropriate system of illumination. Accordingly, the lighting method determines the capability of AOI systems. By using accurate 3D models of soldered SMDs (Surface Mount Devices), new and novel illumination optimization algorithms are described in this paper. Preparation of 3D models, their verification, the implementation of appropriate reflection properties and method for illumination optimization are also discussed.

Keywords: 3D modelling, illumination, image synthesis, optical inspection, optimization, modelling, synthesis, inspection, AOI, quasi-tombstone detection, LEDs, SMDs, AVI, ICT, PWBs, halogen lamp, xenon lamp, metal vapour lamp, Bright field, Dark field, solder joints, RGB ring, MENISCUS, cohesive force, CGS, vector field, EM, BRDF, CGI, Cook-Torrance model, G factor, incident angle, Microsoft DirectX SDK, HLSL, FOI

Rights & PermissionsPrintExport Cite as


Article Details

VOLUME: 2
ISSUE: 3
Year: 2010
Page: [149 - 162]
Pages: 14
DOI: 10.2174/1876402911002030149
Price: $58

Article Metrics

PDF: 2