基于NSCT红外与可见光图像融合算法优化研究

Improved Infrared and Visible Light Image Fusion Algorithm Based on NSCT

为了在一定程度上为后续图像处理提供更为有效的信息,针对红外与可见光图像融合,提出了一种改进的算法。NSCT用于分解红外图像和可见光图像,采用像素特征能量加权融合规则和邻域方差特征信息融合规则得到其低频和高频系数,最后通过逆NSCT进行图像重构得到融合图像。优化后的算法在融合图像清晰度上比Contourlet算法提高了2.22%,在图像信息丰富程度上提高近3.1%。实验结果表明,该算法可以有效地提高图像融合质量。

In order to provide more efficient information for subsequent image processing, an improved algorithm has been proposed for infrared and visible image fusion. Non-subsampled contourlet transform（NSCT） was used to decompose an infrared image and a visible light image. The pixel feature energy weighted fusion rule and the neighborhood variance feature information fusion rule were used to obtain the low-frequency and high-frequency coefficients. Finally, the inverse NSCT was used to reconstruct the image to obtain a fused image. The fused image resolution of the optimized algorithm is 2.22% higher than that of the Contourlet algorithm, and nearly 3.1% better in terms of image information richness. The experimental results demonstrate that this algorithm could effectively improve image fusion quality.