The invention relates to a method for quantifying the level of agglutination of particles in a sample, in particular a biological sample, and notably blood. The biological sample is positioned between a light source and a matrix photodetector. The image acquired by the photodetector is representative of the level of agglutination of the particles in the sample. The light source emits a light wave, the spectral band of which extends an optimum 400 and 600 nm, which constitutes an optimum an excessively low absorption and excessively high absorption, given the thickness of the sample.