平移不变小波在图像融合中的应用

Application of Shift Invariant Wavelet in Image Fusion

二元树复数小波变换(DTCWT)分析图像纹理特征,它具有:近似的移动不变性;在二维变换中具有良好的方向选择性、完备重构性、有限冗余性和计算高效性等优点.DTCWT能得到较好的图像纹理特征和具有更好的鲁棒性.在图像有较小位移情况下,基于DTCWT算法的融合结果要比DWT好得多,融合效果较好.

We analyze texture features of image using the dual tree complex wavelet （DTCWT）, because it is approximate shift invariance, and has good properties in 2D such as ： good directional selectivity, perfect reconstruction, moderate redundancy, efficient computation. DTCWT can get more robust image texture features than other wavelets. With a little shift between images, the fusion result base on DTCWT methods is much better than DWT.