Application of image processing based on multiple filters in an alignment system

Abstract

Abstract Beam alignment depends on CCD real-time image analysis and processing. In order to improve the quality of the alignment, multiple filters are used in far-field and near-field image processings. These multiple filters are constituted of an average filter and a median filter in different connection sequences, so that they can deal with different kinds of noise. To reduce the effect of the unknown nonlinear relationship between motor running steps and deviation pixels, a feasible methodology is offered to improve this phenomenon and a fuzzy algorithm is applied to the motor feedback control process. Because of the fuzzy control it is not necessary to establish an accurate mathematical model, so the impact of the nonlinear relationship will be reduced.