Method of the Road Lines Recognition in the Maps of Digital Material Based on Improvemented BP Neural Network

Maps made of paper mateial are required to be digital, the necessary step is to recognize and trace the road lines in the maps of digital material. But there are so many crossing points of the road lines in the digital map.In order to recognize a complete road line, the tracing direction at the crossing point must be solved at first. According to present roads crossing pointes recognition algorithm has not obtained good effects. This paper describes the calculation principle of road lines recognition using BP neural network in maps of digital material, and method has been improved. It was proved by experiment that the improved algorithm of the crossing point recognition method is highly effective.