Elforgatott és ékezetes térképi megírások automatikus felismerése

Nowadays, maps are mostly stored in digital form. The next step is to vectorize them. Vectorizing maps manually can be a long process; however, we can utilize the computing speed of computers. Computer automation is a developing area which can be used to vectorize maps. In this work, I developed a very simple method to automatically recognize rotated characters and vectorize them, even when they are accented. It is based on a database, in which all the characters and symbols are stored. The method compares the characters on the map with those stored in the database, and returns the database character with the highest match. The result of the process is an output database, where all the recognized characters are stored with image coordinates and rotation angles. The method can connect characters which are close enough to each other according to certain parameters, and can return the word formed by these connected characters.