【作者】 朱嘉祺； 葛玲玉； 吴礼； 彭树生；

【Author】 Zhu Jiaqi;Ge Lingyu;Wu Li;Peng Shusheng;School of Electronic & Optical Engineering, Nanjing University of Sci.& Tech.;

【机构】 南京理工大学电子工程与光电技术学院；

【摘要】 针对极化被动毫米波图像融合问题,基于非下采样轮廓波变换多方向、多尺度分解和平移不变性等优点,提出了一种基于NSCT的极化被动毫米波图像融合方法。首先利用NSCT对极化被动毫米波图像进行分解,对分解后的低频子带和高频子带分别采用区域能量和视觉显著性自适应加权的融合规则,最终实现了极化被动毫米波图像的融合,并通过评价指标对融合结果进行分析。实验结果表明,与原图相比,该方法得到的融合图像具有更好的视觉效果和更优的量化指标。更多还原

【Abstract】 Aiming at the problem of polarized passive millimeter wave image fusion, a polarized passive millimeter wave image fusion method is proposed based on Nonsubsampled Contourlet Transform(NSCT) which has advantages of multi-direction, multi-scale decomposition and translation invariance. The NSCT is initially applied to decompose the polarized passive millimeter wave images into low-frequency subbands and high-frequency subbands, and the regional energy and visual saliency adaptive weighting are employed as the fusion rules for low-frequency and high-frequency subbands respectively to realize the fusion of polarized passive millimeter wave images. The fusion performances are evaluated with several metrics. The experimental results show that the fused image obtained by the proposed method has a better visual and quantitative performance than the original images.更多还原