Abstract: A method for correcting an image obtained by a matrix of photosensitive points finds particular utility when the matrix is subjected to an electromagnetic disturbance. The method of correction comprises a first step of row-by-row reading of the matrix. The signals read during a first reading of each row represent the charges accumulated subsequent to exposure of the matrix to luminous radiation and make it possible to form, for each column of the matrix, a discrete signal. The signals read during a second reading represent the charges accumulated in the absence of exposure and form, for each column of the matrix, a discrete signal. In a second step of the method, a signal corresponding substantially to the signal which would have been formed at the time of the first reading in the absence of exposure is determined. In a third step, the signal is subtracted column by column from the signal.