Color imaging by using of DSP implementation of different filters

摘要

Non linear filtering techniques are presented to suppress impulsive noise and preserve fine image details. Directional processing, a non parametric approach and order statistics filters have been investigated. The previously presented novel algorithm MM-KNN (median M-type k-nearest neighbor) (Gallegos-Funes, F.J. et al., Elec. Lett., vol.38, no.15, p.786-7, 2002) is adapted to color imaging. Different criteria, such as impulsive noise suppression, edge and detail preservation, have been used to characterize the performance of color imaging. The analyzed algorithms have been implemented on a Texas Instruments DSP to obtain the values of processing time when different intensity noise corrupts the color image.