Such methods firstly perform multi-scale analysis on the source image, such as decomposing it into multiple transform domain subbands or decomposition layers with different scales and directions. These decomposition layers reflect image components with different properties in the source image. Then, combine the application scenarios to formulate fusion rules and apply these rules to fuse the corresponding decomposition layers. Finally, the fused sub-bands are transformed into the spatial domain through the inverse transformation of multi-scale analysis to obtain the fused image. With the continuous evolution and development of multi-scale analysis theories such as pyramid transform, wavelet transform, and contourlet transform, their approximation and expression of images are more refined and accurate, so this kind of fusion method can often obtain good fusion quality.