The importance of automated optical inspection (AOI) is growing steadily in electronic technology because of the increasing number and decreasing size of placed surface mounted devices. Lowering the slip-through and pseudo failures is one of the most important aim of developing AOI technology. To achieve this, engineers have to optimalize the inspection parameters, which is still an intuitive, not automated task, and it cannot be guarateed that AOI devices work optimally, meaning taking full advantage of technological possibilities. To make this task easier, one of the important developing directions of AOI is creating 3D device and solder meniscus models with the use of computer generated imagenary (CGI), and by modeling the possible lighting options, choosing the one that makes it the easier to test the existence or soldering of a device or any other AOI task, thus generating the least slip-through and pseudo failure possible.
During working on my thesis, I developed the popular, widely used Cook-Torrance reflectance model. The model calculates the distribution of reflected light from a given surface from physical equations (using the Bidirectional Reflective Distribution Function). The Cook-Torrance model uses two kinds of reflected light types: the diffuse component means the perfectly evenly distributed reflected light, the specular component means the light ray reflected in one exact direction by the rules of geometrical optics. The later is characteristic to metals. To calculate the specular component, the value of the surface roughness is necessary. The model handles the surface as one that consists of small, mirror-like plane sections (called microfacets) that reflect the light perfectly, surface roughness means the root mean square value of the slopes of these microfacets. In my thesis, I have measured the value of the roughness parameter for leaded and lead-free solder alloys, and I have verified my results with software simulation.