Researching of rank detector on grayscale images

We propose an image processing method using a rank algorithm, the advantages of which allow us to a perform data processing task in a-priori uncertainty conditions such as background noise distribution. Thus the proposed method appears to be robust relative to background noise jitter or changing light conditions. In spite of the input image´s dynamic properties, a constant false alarm probability is provided. The goal of our method is to provide high-quality features extraction in the task of object recognition