The Research of Reducing Misregistration Based on Image Mosaicing

Image mosaicing has been collecting widespread attention because it has become as an important tool for several different areas. Among other methods, homograph-based methods are the most accurate in the geometric sense. This is because these methods use planar projective transformation, which considers perspective effects as a geometric transformation model between images. We propose an automatic image mosaicing method which can construct a panoramic image from a collection of digital still images. These methods, however, have a problem of misregistration in the case of general scenes with arbitrary camera motion. Our method has been tested with several image sequences and comparative results are presented to illustrate its performance.