基于改进的拉普拉斯金字塔变换的图像融合方法

介绍了传统的拉普拉斯变换存在的缺点以及改进的拉普拉斯变换的基本原理和重构方法,分析了新的重构算法能够在重构时抑制噪声的特点。提出了在图像融合过程中会引入噪声的问题,并通过实验分析了图像融合过程中引入的噪声情况。使用新的重构算法能够在图像重构过程中,而不是在系数处理过程中,有效抑制融合噪声。给出了基于改进的拉普拉斯变换方法进行图像融合的基本架构,并对变换系数的设置与融合过程的处理进行了详细的介绍。仿真实验的主客观性能比较表明,基于改进的拉普拉斯变换的图像融合方法比其它几种基于金字塔变换的融合效果要好得多。

Multi-Sensor Image Fusion Based on Improved Laplacian Pyramid Transform

The shortcoming of the traditional Laplacian pyramid(LP) transform is pointed out,and the principle and reconstruction of the improved Laplacian pyramid transform are introduced.The new reconstruction method can restrain the noise produced in the reconstruction process.The noise in the processing of the image fusion is analyzed theoretically and experimentally.The new reconstruction method can restrain the noise in the reconstruction process rather than in the index processing produre.The image fusion framework using the improved Laplacian pyramid transform is proposed and the way to determine the coefficients and the fusion procedure are introduced in detail.The subjective and objective indexes of simulation show that the proposed algorithm works better in preserving the edge and texture information than other Laplacian pyramid methods.