Methods and apparatuses for identifying objects in a digital image include a thresholding unit and a threshold calculation unit. The thresholding unit is designed to receive the digital image and compare the luminance values of the digital image with threshold values of a threshold matrix in order to create a binary image on the basis of the comparison. The threshold calculation unit is designed to estimate, on the basis of a first and second division of a reference image which corresponds to the digital image into first and second subareas, a background luminance value for each first subarea and an object luminance value for each second subarea. The apparatuses and the methods permit quick and memory-efficient identification of the individual objects, even if the digital image contains variation in the luminance and/or sharpness.