Multisensor concealed weapon detection by using a multiresolution mosaic approach

Image mosaicing is a technique to combine two or more images or image regions into a new one with an invisible seam. Images acquired by heterogeneous image sensors may provide complementary information about the scene, and a so-called multiresolution image fusion technique is applied to integrate the information at the pixel level. However, when the images are significantly different, consider a visual and an IR image for example, the performance of multiresolution (MR) fusion algorithms is not always satisfactory. When the application of concealed weapon detection is considered, an alternative solution is to detect the concealed weapon in an IR or a mm wave image first, and then fuse/embed this region in the visual image. In this study, the MR mosaic technique is employed to combine the visual and IR or mm wave images, where the concealed weapon is first detected by a fuzzy k-means clustering method from the IR or mm wave image. The synthesized image retains a quality comparable to the visual image while the region of the concealed weapon is highlighted. Experimental results indicate the efficiency of the proposed approach.