Image Alignment and Stitching: A Tutorial

← Back to document page

Image Alignment and Stitching: A Tutorial1Richard SzeliskiLast updated, December 10, 2006Technical ReportMSR-TR-2004-92This Tutorial reviews Image Alignment and Image stitching algorithms. Image align-ment algorithms can discover the correspondence relationships among images withvarying degrees of overlap. They are ideally suited for applications such as videostabilization, summarization, and the creation of panoramic mosaics. Image stitch-ing algorithms take the Alignment estimates produced by such registration algorithmsand blend the images in a seamless manner, taking care to dealwith potential prob-lems such as blurring or ghosting caused by parallax and scene movement as well asvarying Image exposures.

Dec 10, 2006 · Name Matrix # D.O.F. Preserves: Icon translation h I t i 2×3 2 orientation +··· rigid (Euclidean) h R t i 2×3 3 lengths +··· similarity h sR t i 2×3 4 angles +··· affine h A i 2×3 6 parallelism +··· projective h H˜ i 3×3 8 straight lines Table 1: Hierarchy of 2D coordinate transformations. The 2 ×3 matrices are extended with ...

  Image, Transformation, Rigid, Alignment, Stitching, Image alignment and stitching