Abstract <p>This paper describes an approach to image vectorization based on complex morphological processing of vector data using vector morphological operations on paths (lines tracing the contours of the depicted objects). A review of existing vectorization methods, both algorithmic and machine learning based, is conducted. A mathematical representation of a path is described, including various path characteristics, path structures, and path decomposition into segments. A table of blocks and types of vector morphological operations is provided, and simple operations included in the path correction procedure are described. An experiment was conducted on image vectorization, which included, among other things, a procedure for filtering vector information (removal of noise path structures).</p>