Algorithms for pitch distance determination

Abstract

The paper describes algorithms used for the determination of pitch distances in the calibration of optical scales. The algorithms are divided into two groups: the non-contextual and the contextual algorithms. The first group is used to calculate the position of the edges in the digital image captured by the camera is either based on the gradient or the so called Laplacian method without any prior knowledge of the shape of the object. The contextual group of algorithm make use of the of the a priori knowledge of the edge or shape. The information carried by gray level values of the pixels is used to reach subpixel resolution. As a result pixels are categorized as black, white or edge pixels. The algorithms in the second group calculate the centreline of the graduation or measuring line either using Gabor transformation or from the centre of gravity formed by the gray values of the individual pixels. The resistance of the algorithms to various distortions on the scale was investigated using real objects.