Improved image mosaicing technique using marker controlled watershed segmentation

Seamless mosaic generation is a challenging issue in the field of image processing. Multi date and multi time images vary in terms of radiometry as well as geometry of viewing, this makes the task at hand more intricate. These variations are taken care of with the help of radiometrie normalization and geo-correction. In this paper an approach is proposed for mosaicing geo corrected images acquired on different days. This approach uses inter scene normalization of the images and utilizes a mathematical morphological operator to find out the best possible seam line from the overlap area. Overall visual quality of the mosaiced product is analyzed to verify the algorithm.