基于Surfacelet变换的井下图像融合技术研究

Underground Image Fusion Based on Surfacelet Transform

首次将Surfacelet变换算法应用于井下图像融合。通过多方向分解和树结构滤波器组(3D-DFB)实现图像的不同尺度、不同频率方向的分解。对高低频子带系数采用不同的加权规则进行融合。既保留了Surfacelet变换的多分辨率特性,又具有平移不变性。实验结果表明,相对于传统的变换法,取得了更佳的融合效果。

The Surfacelct transform is used to fuse the underground image for the first time.The original image is decomposed in direcfions of different scales and freguencies by muiti-direclional decomposition and tree filter banks（3D-DFB）.The obtained Low and high freguency coefficients are fused by different weighted regulations The algorithm keeps the multiple distinguishability of Surfacelet transform and has the unvarying validity during moving.The experimeafal results show that the fusion effect of the algorithm is better than that of the traditional methods.