Image fusion is a process of generating an informative image from more than one complementary image. It finds applications in military, navigation, concealed weapon detection, medical imaging, digital photography and remote sensing etc. A new image fusion method based on two-scale image decomposition and saliency map detection is proposed in this paper for multi-sensor images. The algorithm is as follows: First, each source image is decomposed into base and detail layers. Second, saliency map of each source image is calculated with help of frequency tuned saliency map detection. Third, detail images are fused by using the proposed decision map based on the saliency maps and base layers are averaged to get the fused base layer. Finally, fused image is generated by taking the linear combination of fused base and detail layers. This algorithm is very advantageous because the saliency map used in this paper highlights the saliency information uniformly with well defined boundaries. So the decision map based on these saliency maps can effectively transfer the complementary information from source images to the fused image. Unlike traditional multi-scale decomposition fusion methods, proposed method uses two-scale decomposition to get base and detail layers. So it is computationally efficient. Outcomes of the proposed method are compared with existing multi-scale decomposition techniques along with spatial domain techniques with help of traditional and objective fusion metrics. Results reveal that proposed method outperforms the existing methods.