Neural nets for aligning optical components in harsh environments: Beam smoothing spatial filter as an example

摘要

The goal is to develop an approach to automating the alignment and adjustment of optical measurement, visualization, inspection, and control systems. Classical controls, expert systems, and neural networks are three approaches to automating the alignment of an optical system. Neural networks were chosen for this project and the judgements that led to this decision are presented. Neural networks were used to automate the alignment of the ubiquitous laser-beam-smoothing spatial filter. The results and future plans of the project are presented.