A Robust Image Mosaicing Using Improved SIFT Technique

Image mosaicing is a technique that combines fragmented images to produce a more comprehensive depiction of a landscape or situation. This process has numerous applications in various fields, including motion detection, pixel density improvement, land surface surveillance, and diagnostic imaging. Over the past two decades, researchers have developed several algorithms and techniques for mosaicing images. A deep learning approach is used in this research study to present an innovative method for flawlessly mosaicing photographs. In this approach, the data collection unit for the mosaic picture is a mobile device, such as a smartphone. In order to extract descriptors from the photographs, we applied a modified version of the scale-invariant feature transform (SIFT) methodology. In order to determine the nature of the transformation that occurred between the images, the homography matrix was constructed by making use of the RANSAC method. After that, the input images were homographically aligned by utilising the computed homography matrix. The final product of applying this strategy will be a clear mosaic comprising various components of a variety of scenes, which will be assembled into a comprehensive image. The suggested method provides robustness against changes in lighting conditions and scene alterations by making use of SIFT. These changes can occur at any time. In addition to handling picture rotations, translations, and scaling, the method can also deal with image scaling. The accuracy of the study was determined to be 84.32%.