| 摘要 |
<p>The invention relates to matrix image sensors, and more particularly to a method for correcting the spatial noise engendered by the dispersion of the physical properties of the various pixels of the matrix. According to the invention, a signal Xi(L) corresponding to the illumination of a pixel Pi by a luminance L is established pixel by pixel; a fixed reference K0 is defined for this signal, the value K0 being chosen such that for a given luminance L0 the pixels all provide an identical signal Xi(L0) equal to K0; and a corrected signal X*i(L) is established by applying to the signal Xi(L)-K0 thus referenced a multiplicative corrector coefficient specific to the pixel considered, the corrector coefficient being determined on the basis of the signal Xi(Lr) provided by the pixel Pi for a determined reference luminance Lr which is the same for all the pixels. The corrector coefficient is preferably stored in an initialization register REG1 specific to the pixel, which contains a value dTi of modulation of integration time of a counting frequency representing the luminance.</p> |
