The marine and air background images obtained by a single shipborne infrared sensor often have problems such as low contrast between target and background, high noise, and lack of complete target details, which bring great difficulty to the extraction of ship targets. This paper analyzes the basic features of the ship target in medium/long wave infrared image and proposes the basic model of ship target extraction based on the Markov Random Field (MRF) theory. According to the two-band target and background prior probability distribution, the energy minimization framework is added. Regional believable propagation (BP) algorithm is used to perform global optimization of the model, and image segmentation label is estimated according to MAP criteria. The experimental results show that the fusion extraction algorithm can retain the effective components in the original dual-band infrared image, and the extraction efficiency and accuracy are higher.

Date Published: 12 March 2019
PDF: 9 pages
Proc. SPIE 11023, Fifth Symposium on Novel Optoelectronic Detection Technology and Application, 110231V (12 March 2019); doi: 10.1117/12.2521902

Published in SPIE Proceedings Vol. 11023:
Fifth Symposium on Novel Optoelectronic Detection Technology and Application
Qifeng Yu; Wei Huang; You He, Editor(s)