摘要
为了提高低可见度情况下红外与可见光图像融合质量,提出了一种基于有限离散剪切波变换(FDST)和图像显著性检测算法(MSS)的融合方法。利用显著性检测算法对红外和可见光图像进行检测。采用有限离散剪切波变换对红外和可见光图像进行分解。根据显著性检测结果中亮度越高的区域属于显著性目标这一特点,指导低频子带图像融合。高频子带图像采用"绝对值和取大"的融合策略。实验结果表明提出的融合方法在主观视觉效果和客观性能指标上均取得了较好的效果。
Aiming at improving the quality of infrared and visible images fusion in low visibility condition, a fusion method based on FDST（finite discrete shearlet transform） and MSS （saliency detection using maximum symmetric surround） was proposed.Using MSS for detecting infrared and visible images, using FDST, infrared and visible image can be decomposed.In detection results, higher luminance region belong to saliency region, this characteristic can be used to guide low-frequency sub-band images fusion.High-frequency sub-band images were fused with absolute value sum maximum selectivity.Fusion results indicate that the proposed method is performed very well in the subjective visual and objective evaluation.
出处
《科学技术与工程》
北大核心
2017年第6期215-219,共5页
Science Technology and Engineering
基金
成都市科技惠民项目(2015-HM01-00293-SF)资助
关键词
有限离散剪切波变换
显著性检测
图像融合
区域显著值和取大
finite discrete shearlet transform saliency detection using maximum symmetric surround image fusion regional significant value sum
