Soft Mathematical Morphology For Binary Images Filtering

The paper proposes soft morphological operators and algorithms to remove noise from binary images. The features of the algorithm are the presence of additional parameters, such as the threshold of filtering and number of iterations, and the ability to remove noise: islands, pin holes, mousebites and spurs in one pass. Application of these algorithms can bring less distortion during operation, even if the image objects are spaced one pixel apart. There are results of the experiments on the example of the PCB artwork.