Image segmentation algorithm based on Markov random field(MRF) for radiography

Boundary of different materials is one of the focus in flash X-ray radiography, low signal-to-noise ratio and microcosmic structure block the exact distilling of material boundary. In this paper, Markov random field method is researched, Markov model is used to describe the relativity between different pixels, to reduce the number of false target root in noise; through the variance of pixels in template with hollow center the energy cliques function is described, the segment result of microcosmic structure and segment precision of flash X-ray radiography are improved. Simulation shows the modified MRF method in this paper can get better segment result.