AUTOMATIC RECOGNITION OF CIVIL INFRASTRUCTURE OBJECTS IN MOBILE MAPPING IMAGERY USING A MARKOV RANDOM FIELD MODEL
Zhuowen Tu, Ron Li · 2000
Information technology is increasingly used to support civil infrastructure systems that are large, complex heterogeneous, and distributed. These dynamic systems include communication systems, roads, bridges, traffic control facilities, and facilities for the distribution of water, gas and electricity. Mobile mapping is a new technology to capture georeferenced data. It is, however, still not practical to extract spatial and attribute information of infrastructure objects fully automatically. In this article, a framework for 3D-object recognition is proposed according to a viewpoint dependent theory. A novel system that generates hot-spot maps using color indexing and edge gradient indexing and recognizes traffic lights using MCMC (Markov Chain Monte Carlo) method is proposed. The hot-spot map generation method we developed is much faster than general color image segmentation and thus is practical to be applied in a recognition system. In this approach, both top-down and bottom-up methods are combined by the MCMC engine, which not only recognizes traffic lights but also tells us their poses. This system is robust for different degrees of illumination and rotation. 1