The problem of raster images vectorization was analyzed and researched in this work. Existing vectorization systems are quite expensive, the results are inaccurate, and the manual vectorization of a large number of drafts is impossible. That‘s why our goal was to design and develop a new raster images vectorization system using our suggested automatic vectorization algorithm and the way to record results in a new universal vectorial file format. The work consists of these main parts: analysis, design, user documentation and experimental research. Conclusions of the analysis of this field are that the optimal solution of vectorization is the synthesis of automatic and manual vectorization. Users’ requirements were documented and raster image vectorization system designed. It was noticed that the result of vectorization depends on: vectorization algorithm; vectorization parameters; quality of raster image and user experience in this field. The core problems of vectorization are: symbol, intersection and corners recognition, lines‘ and their endings‘ types, filled areas. Vectorization results of our system and Corel TRACE were compared during the experimental research.