在二代Curvelet变换理论的基础上, 结合红外和可见光图像的成像源理以及物理特征, 为了更加有效地对图像像素点中信息质量进行测量, 提出了一种二代Curvelet算法用于红外与灰度可见光图像融合。算法中将第二代Curvelet变换域引入活度加权以对图像中像素点信息质量进行评定, 并将方向对比度以及活度加权对显著水平进行测量, 提出了用显著水平测量的方法对系数进行选择, 进而提取红外图像的目标特征以及灰度可见光图像中大量的背景特征信息。仿真结果表明, 该算法能够在保留原图像中重要信息的同时, 融合后的图像边缘细节更加清晰、边缘更加平滑。

In this paper, based on the second generation based on the Curvelet transform theory, combined with infrared and visible light imaging and physical characteristics, in order to more effectively measure information quality in the pixels of the image, a second-generation Curvelet algorithm is proposed for IR fused with grayscale optical images. In the algorithm introduced the second generation Curvelet transform domain activity weighted to the pixels in the image information to assess their quality and directional contrast and weighted to measure the significant level, presents significant measurement coefficient with the method of selection and extraction of infrared image of the target characteristics, and gray-scale characteristics of visible in the image a lot of background information.. Simulation results show that the algorithm can retain important information in the source image at the same time, edge details of the image much clearer and edge smoother after fusion.