Book contents
- Frontmatter
- Contents
- List of contributors
- Foreword by Jón A. Benediktsson
- Acknowledgements
- PART I The Importance of Image Registration for Remote Sensing
- PART II Similarity Metrics for Image Registration
- PART III Feature Matching and Strategies for Image Registration
- 7 Registration of multiview images
- 8 New approaches to robust, point-based image registration
- 9 Condition theory for image registration and post-registration error estimation
- 10 Feature-based image to image registration
- 11 On the use of wavelets for image registration
- 12 Gradient descent approaches to image registration
- 13 Bounding the performance of image registration
- PART IV Applications and Operational Systems
- PART V Conclusion
- Index
- Plate section
- Plate section
- References
10 - Feature-based image to image registration
from PART III - Feature Matching and Strategies for Image Registration
Published online by Cambridge University Press: 03 May 2011
- Frontmatter
- Contents
- List of contributors
- Foreword by Jón A. Benediktsson
- Acknowledgements
- PART I The Importance of Image Registration for Remote Sensing
- PART II Similarity Metrics for Image Registration
- PART III Feature Matching and Strategies for Image Registration
- 7 Registration of multiview images
- 8 New approaches to robust, point-based image registration
- 9 Condition theory for image registration and post-registration error estimation
- 10 Feature-based image to image registration
- 11 On the use of wavelets for image registration
- 12 Gradient descent approaches to image registration
- 13 Bounding the performance of image registration
- PART IV Applications and Operational Systems
- PART V Conclusion
- Index
- Plate section
- Plate section
- References
Summary
Abstract
Recent advances in computer vision address the problem of registration of multiple images or entire video sequences. Such registration methods have a wide variety of application in constructing mosaics, video summarization, site modeling and as preprocessing for tasks such as object tracking and recognition. In this chapter we present a variety of registration techniques that utilize image features such as points and contours. Computational issues such as robustness to data outliers and recent developments in accurate feature extraction are discussed. A correspondenceless method that works on multimodal images is outlined. We also present approaches that efficiently utilize the information redundancy in a sequence of images to solve the problem of image registration. All of these methods are illustrated with appropriate examples.
Introduction
The underlying geometry of image formation has been well studied over the recent years in the discipline of computer vision (Hartley and Zisserman, 2004). This understanding of the image geometry has been accompanied by increasingly sophisticated computational models that can be solved on modern hardware. Many methods and ideas developed for solving various aspects of the motion estimation problem in computer vision are applicable to problems relating to image registration. In particular, image features like points, edges, and contours have been used in a range of applications like the construction of mosaics from video sequences, shape estimation, object tracking and recognition, etc. In this chapter, we describe a variety of methods dedicated to utilizing image features for solving the problem of registration.
- Type
- Chapter
- Information
- Image Registration for Remote Sensing , pp. 215 - 239Publisher: Cambridge University PressPrint publication year: 2011