基于改进Tetrolet变换的图像融合算法研究

针对红外图像目标物体能量高以及可见光图像细节信息丰富的特点,提出一种基于改进的Tetrolet变换的红外与可见光图像融合算法.对红外和可见光图像进行多尺度、多方向分解,在低频融合规则上,对区域能量进行适当缩放,突出红外目标,保留可见光背景信息.实验结果表明,对Tetrolet变换模板的选择的改进,有助于获取更多高频信息;融合算法相对于传统的算法不仅增强了图像对比度,改善了主观视觉效果,而且在客观标准上有了一定提高.

Research of Image Fusion Algorithm Based on Improved Tetrolet Transform

An improved tetrolet transform used in the fusion of infrared image and visible image was proposed according to high-energy of the target object of infrared image and the rich detail information of visible image.The images are decomposed at multi-scales and multi-directions,and in the fusion rule of low-frequency,appropriate scaling of regional ener-gy is used to prominent infrared target and retain background information of visible image.The experimental results show that more high-frequency information is retained by the improved selection of tetrolet transform template. Compared to the traditional fusion algorithm,the proposed fusion algorithm enhances the image contrast,improves subjective visual effects,and also raises the objective criteria.