Adaptive Filtering Algorithm for Vision Inspecting of Color Printing
Yue Mei · Yibiao jishu yu chuanganqi · 2010
Impulse noise in detected image can severely affect subsequent image processing;therefore impulse noise filtering is an essential part of color printing vision inspection algorithms.An adaptive filter composing of two cascaded subunits was presented.The first subunit is a group of sigma vector median filters,and each of them has a different switching threshold.The second subunit is a first-order Sugeno-type Neuro-fuzzy network whose parameter is automatically tuned and fixed by learning to meet the complicated nonlinear mapping relation,and the optimized result is finally obtained.Experimental results show the proposed filter is able to efficiently suppress the impulse noise without distort image details.