Detection of peaks nanoscale images in noises

A new noise-immune algorithm for estimating the position and intensity of the peaks of nanoobjects is proposed. The method is based on image processing by columns and rows using the algorithm of three-point extremes search in a sliding data window. The method allows to estimate the position and intensity of peaks with a signal-to-noise ratio up to three with probability of detection up to unity, zero probability of peak miss and false alarm.