Research on Road Marking Image Recognition Method Based on Deep Belief Network

Neural network is a key part of image recognition and classification. It is also one of the hotspots in recent years. Deep belief network, as a method of machine learning, has strong generalization and gets more and more attention from experts and scholars at home and abroad. Aiming at these two research hotspots, a method of road marking image recognition based on deep belief network (DBN) is proposed in this paper. Four experiments below show that it has high accuracy and practicability, which provides an important theoretical basis for highway’s construction and maintenance in special areas.