The adaptive process for the removal of ghost images in video signals

摘要

An adaptive method for suppressing video signal x(k) echoes in equalizers of TV sets including digital filters whose coefficients are updated in an adaptive and iterative manner using an LMS (Least Mean Square) algorithm until the difference, or output error, between a target output signal d(k), called the reference signal, and an outgoing signal y(k) from the equalizer is gradually reduced, comprises the steps of: applying a "combing" technique to an original filter having K*N coefficients in order to select K comb filters having N coefficients each;applying said LMS algorithm, with a variable convergence factor (µ) to each individual comb filter for a predetermined number of iterations;gathering the resultant configurations of the comb filter coefficients and selecting a subfilter with N largest modulo coefficients therefrom;updating the values of said N coefficients again by reiterating the LMS algorithm to said subfilter for a limited number of iterations;zeroing all the coefficients with a lower modulo than a predetermined threshold;selecting a group of F coefficients by a slotting operation across those of said coefficients which have a cluster value: andupdating the value of said group of F coefficients, by reiterating the LMS algorithm, until the output error e(k) becomes smaller than a predetermined value.