Image deformation in images stitching

• Main Author: Goh, Ju Han.
• Format: Final Year Project Report
• Language: English; English
• Published: Universiti Malaysia Sarawak (UNIMAS) 2017
• Subjects: T Technology (General); TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://ir.unimas.my/id/eprint/25629/1/Image%20deformation%20in%20images%2024pgs.pdf; http://ir.unimas.my/id/eprint/25629/4/Goh%20Ju%20Han.pdf; http://ir.unimas.my/id/eprint/25629/

Panorama image stitching is a process of combining two or more images with overlapping fields of view to produce a 360ºpanorama image. This technique is widely used in medical imaging, multiple image super-resolution, video stitching and object insertion. A stitched image can be presented in several types of views such as cylindrical, rectilinear, spherical and stereographic. The projections of the image give a distortion result where the object in the image is curved or bent to fulfil the requested projection. Furthermore, the curved or bent image might lose information from the original images. Therefore, flattening the curved area can help to restore the lost information and present the image in a better view. In this research, the flattening methods are normally used in developable surface but less in image processing fields. Moreover, the research materials on this topic are limited due to the new concept of flattening methods in image processing. Additionally, both concept of image flattening and image stitching is combined to perform a new method on image deformation. In this project, the aim is to develop an image deformation algorithm to flatten the curved image in panorama view. Furthermore, an image stitching algorithm is used to stitch the deformed images into a plane image. Moreover, the experiment result showed that the method is flexibly to deform different projection of image and able to perform a good result.