Linear N-Point Camera Pose Determination
IEEE Transactions on Pattern Analysis and Machine Intelligence
Multiple view geometry in computer visiond
Multiple view geometry in computer visiond
Elements of Photogrammetry(with Applications in GIS)
Elements of Photogrammetry(with Applications in GIS)
Graphcut textures: image and video synthesis using graph cuts
ACM SIGGRAPH 2003 Papers
Unsupervised 3D Object Recognition and Reconstruction in Unordered Datasets
3DIM '05 Proceedings of the Fifth International Conference on 3-D Digital Imaging and Modeling
Multi-Image Matching Using Multi-Scale Oriented Patches
CVPR '05 Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05) - Volume 1 - Volume 01
MapCruncher: integrating the world's geographic information
ACM SIGOPS Operating Systems Review - Systems work at Microsoft Research
Automatic Panoramic Image Stitching using Invariant Features
International Journal of Computer Vision
ORIENT-CAM, a camera that knows its orientation and some applications
CIARP'06 Proceedings of the 11th Iberoamerican conference on Progress in Pattern Recognition, Image Analysis and Applications
| Hi-index | 0.00 |
Commercial aerial imagery websites, such as Google Maps, MapQuest, Microsoft Virtual Earth, and Yahoo! Maps, provide high- seamless orthographic imagery for many populated areas, employing sophisticated equipment and proprietary image postprocessing pipelines. There are many areas of the world with poor coverage where locals might benefit from recent, high-resolution orthographic imagery, but which do not fit into the schedules and scaling model of the big sites. This paper describes MapStitcher, a system that orthorectifies and geographically registers imagery using only low-cost capturing equipment. MapStitcher combines manually-entered relationships between images and known ground references with a MOPs-based image-stitching technique that automatically discovers image-to-image relationships. Our image registration pipeline first extracts and matches feature points, then clusters images, then uses RANSAC-initialized bundle adjustment to simultaneously optimize all constraints over the entire image set. Simultaneous optimization balances the requirements of precise stitching and absolute placement accuracy. We used this technique to image a portion of the Skagit River Valley in the vicinity of the town of Concrete, WA (pop. 790) at 0.15 m/pixel. Our technique is more accurate than stitching followed by "rubber-sheeting" (deforming the stitched image into global coordinates), while it also requires less effort and produces a better-stitched composite than rubber-sheeting images separately.