Edge-preserving wavelet-based multisensor image fusion approach

Images emanating from multiple sensors have been successfully exploited to reduce human and machine errors in practical vision systems. Multiresolution-based schemes have shown interesting potential in the fusion of images obtained from possibly different types of sensors that need to be combined. However, most of the proposed schemes treat all image features equally regardless of their local importance. On the other hand, the human visual system is more sensitive to edges and sharp details. We propose an image fusion scheme where image edges, characterized by wavelet maxima, are considered separately from plain and low activity image regions. This edge-guided fusion offers a trade-off between feature-based and pixel-level fusion schemes. Images are combined in the wavelet domain using a multiresolution representation that is more sensitive to image edges. A comparison of the proposed method with current multiresolution-based fusion schemes shows that the proposed method can achieve better performance in combining and preserving important details in the combined images.