Abstract: A focussing system in which a detection and processing device (2) scans the image plane, in raster fashion, to generate a video signal representing the intensity of radiation prevailing at successively scanned locations. A selection circuit (6) contains a programmable memory which generates addresses suitable for selecting discrete sets (S) of neighbouring locations and a circuit (7) is responsive to the addresses to access to a processing circuit (8) only those portions of the video signal which correspond to the sets (S) of locations. The processing circuit (8) generates difference signals representing differences of intensity between neighbouring locations and evaluates the number (P.sub.tm) of difference signals exceeding each of a number of preset threshold values (t.sub.m) to generate a histogram. The numbers constituting the histogram are then weighted and scanned to generate a control signal (F) representing the focus.

Abstract: A set of stores contains reference information representing a reference pattern. A scene is scanned and information derived from each of a succession of portions of the scene is fed to the stores which produce correlation signals if the scene information of that portion correlates with the reference information.The position of the portion of the scene which is fed to the stores is tracked, and the position associated with the highest correlation score is used as a representation of the position in the scene of the reference pattern.

Abstract: A digital buffer store 25 (FIG. 2) stores digitized samples of a video signal. A circuit 206 determines the mean grey level (f(b)) of the stored signals and a circuit 28 forms a histogram indicating the number of pixels having various ones of 8 possible grey leves. A circuit 29 determines the dispersion of the histogram about the mean grey level. In principle, circuit 29 could determine from the means level and the histogram the statistical distribution of pixels amongst the grey levels and from that determine the standard deviation as a measure of dispersion. In practice, the circuit 29 determines an emirical approximation to the dispersion, ##EQU1## where n=number of binary bits=3N.sub.i =number of pixels having a grey level iA=f(b)=mean grey levelN.sub.A =number of pixels having the means grey level.This equation is advantageous as it operates on powers of 2.The circuit of FIG. 2 compares the determined values of f(a) and f(b) with empirically derived desired values and servoes the gain and D.C.

Abstract: When correlating a reference representation with current representations the scales of which are changing, errors occur. In order to reduce the errors, a normal current representation is stored in a store (7), and a scaled current representation is stored in a store (8). Both representations are correlated in a correlator (11) with a normal reference representation stored in a store (9). The reference representation is updated by a control processor unit 5 when the correlation of it with the current representation reaches a peak. Instead of using two current representations a single current representation could be correlated with normal and scaled reference representations.