Infrared target segmentation method with multiscale MRF based on data distribution deduced from KLFCV

Kun Wei, Yong Qiang Zhao, Yong Mei Cheng, Shi Bo Gao, Hong Cai Zhang

科研成果: 期刊稿件文章同行评审

摘要

A new infrared target segmentation method based on infrared multiband images is proposed. First, 8-14μm waveband is subdivided into four bands by light filter, and the single band image is substituted by multi-band's order to increase the image information. Then, based on the KLFCV fuzzy chuster model, a new data distribution used as likelihood distribution of multi-scale MRF is presented. Finally, the corresponding parameter estimation is given by using EM algorithm. From the result of the experiments on artifact data and real scene image segmentation and comparison with the traditional method, the proposed method displays certain improvement in segmentation integrity and noise reduction.

源语言英语
页(从-至)662-667
页数6
期刊Guangdianzi Jiguang/Journal of Optoelectronics Laser
20
5
出版状态已出版 - 5月 2009

指纹

探究 'Infrared target segmentation method with multiscale MRF based on data distribution deduced from KLFCV' 的科研主题。它们共同构成独一无二的指纹。

引用此