Research On Digital Image Fusion Based On Wavelet Transform

Wavelet transfer is a time-frequency analysis method that has developed in the last twenty years. It is a powerful tool to represent the signals in different levels and resolutions. Now wavelet is wildly used in the fields of signal processing, image processing, pattern recognition, machine vision and so on.Image fusion is a combination of two or more different images of the same scene to obtain a new image by using certain algorithms. The fused image contains the complimentary information from the different images so that it is more suitable for segmentation, feature extraction, object recognition and human visual system. Generally, the fusion can be devied into three levels: pixel-leve, feature-level, and decision-level fusion. The pixel-leve fusion is the lowest form of fusion. And it is the foundation of the other two levels of image fusion. This paper is deal with the application of wavelet transform in pixel-level image fusion. We first described the theory of wavelet transform, multiresolution analysis (MRA), and the Mallat algorithms for two-dimensional image decomposition and construction. And then we briefly introduced the concept of image fusion, the level of image fusion, and the method for pixel-level image fusion. The subjective and objective evaluation criterions for the performance evaluation of image fusion are discussed. Next, the choice of the decomposition level and the fusion ruler for the Wavelet based image fusion method are discussed. Following, the influence of four color space transform on the imager quality are compared and discussed. Aimed at the IR and visible image fusion, a color image fusion algorithm using Intensity-Hue-Saturation (IHS) transform and wavelet-based contourlet transform (WBCT) is proposed. The WBCT can offer a rich set of directions and shapes because that it is the combination of the multiscale Wavelet decomposition and the multidirection Contourlet decomposition. Thus, this algorithm is effective in capturing geometric structures in images. The experimental results show that the fused images produced by the proposed method is of better quality than that obtained by the convention method based on the IHS and Wavelet transform.