Research of Image Transform Strategy in Fusion System
In the image fusion system, there is a serious memory hierarchy bottleneck. Thus an important transform strategy is the memory management. Motivated by the characteristics of lifting wavelet algorithm, this paper introduces a linearity-coding transform strategy into the image fusion system, which linearly codes registered input images in advance. Then, the coded pixels are transformed into wavelet coefficients by a proposed transform algorithm-LFWPT. Once completing transform,portions of coefficients immediately participate in fusion computation regardless of other ones whether completing transform or not. At last, fused coefficients are reconstructed to image by inverse LFWPA. The advantages of the proposed transform strategy are not only improving the memory management, but also presenting the in-place computation and parallel-processing of lifting scheme in the image fusion system.Experiment results demonstrate the proposed scheme is more effective and fast than the ones with traditional transform strategies, and well fit for hardware implementation.
image fusion lifting scheme linearity-coding.
SUN Fuxiong
Institute of Information Zhongnan University of Economics and law Wuhan, Hubei Province 430074, China
国际会议
武汉
英文
660-665
2007-07-25（万方平台首次上网日期，不代表论文的发表时间）