Fusion of Visible and Infrared Image Based on Stationary Tetrolet Transform

Yu Huang,Dexing Zhang,Baohong Yuan,Jingzhong Kang

DOI: https://doi.org/10.1109/yac.2017.7967529

2017-01-01

Abstract:Tetrolet transform has a better directionality of the structure and can express texture features of image precisely in dealing with high dimensional signal. But, tetrolet decomposition algorithm is easy to cause blocking artifacts in images fusion. In order to reduce the blocking artifacts resulted from tetrolet transform algorithm, a new image fusion technique based on stationary tetrolet transform is proposed. Firstly, the visible image and infrared image can be decomposed into low-frequency and high-frequency coefficients on different scales and in various directions using stationary tetrolet transform. For the low-frequency coefficients, the local region gradient information was applied to get the low-pass fusion coefficients. For the each directional high frequency coefficients, the larger region edge information measurement factor is used to select the better coefficients for fusion. Finally, the inverse stationary tetrolet transform was utilized to obtain fused image. Experimental results show that the proposed algorithm gives more satisfactory results than the traditional image fusion algorithms in preserving the edges and texture information.