Analysis of moving DLT, image and seam selections algorithms with MS ICE, autostitch, and OpenCV stitcher for image stitching applications

The objective of image stitching is to merge several images into a single seamless mosaicked image. Image stitching finds its applications in numerous fields like image stabilization, medical imaging, satellite imaging etc. The results of the existing algorithms are unsatisfactory in specific images, where the images do not differ purely by rotation. Even, the processing time is excessive if there are numerous images with similar scenes or with many blurred images. This paper analyses two efficient image stitching algorithms, a) Moving Direct Linear Transformation (Moving DLT) and b) image and seam selections. In addition, we describe how each of these algorithms incorporate stitching multiple images to create big panoramas using techniques like bundle adjustment and stitching label estimation using particle filtering with partitioned sampling. The experimental results are compared with MS ICE, Autostitch, and OpenCV stitcher.

[1]  Richard Szeliski,et al.  Construction of Panoramic Image Mosaics with Global and Local Alignment , 2001 .

[2]  Cheng-Ming Huang,et al.  Efficient Image Stitching of Continuous Image Sequence With Image and Seam Selections , 2015, IEEE Sensors Journal.

[3]  Michael S. Brown,et al.  Constructing image panoramas using dual-homography warping , 2011, CVPR 2011.

[4]  Richard Szeliski,et al.  Systems and Experiment Paper: Construction of Panoramic Image Mosaics with Global and Local Alignment , 2000, International Journal of Computer Vision.

[5]  Yoichi Sato,et al.  Shape-Preserving Half-Projective Warps for Image Stitching , 2014, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

[6]  David L. Milgram,et al.  Computer Methods for Creating Photomosaics , 1975, IEEE Transactions on Computers.

[7]  Shmuel Peleg,et al.  Elimination of seams from photomosaics , 1981, Computer Graphics and Image Processing.

[8]  Andrew W. Fitzgibbon,et al.  Bundle Adjustment - A Modern Synthesis , 1999, Workshop on Vision Algorithms.

[9]  Richard Szeliski,et al.  Recovering 3D Shape and Motion from Image Streams Using Nonlinear Least Squares , 1994, J. Vis. Commun. Image Represent..

[10]  Scott Schaefer,et al.  Image deformation using moving least squares , 2006, ACM Trans. Graph..

[11]  Michael S. Brown,et al.  As-Projective-As-Possible Image Stitching with Moving DLT , 2013, IEEE Transactions on Pattern Analysis and Machine Intelligence.