Download presentation
Presentation is loading. Please wait.
Published byMelanie West Modified over 9 years ago
1
Image Stitching Shangliang Jiang Kate Harrison
2
What is image stitching?
4
Introduction Are you getting the whole picture? –Compact Camera FOV = 50 x 35°
5
Introduction Are you getting the whole picture? –Compact Camera FOV = 50 x 35° –Human FOV = 200 x 135°
6
Introduction Are you getting the whole picture? –Compact Camera FOV = 50 x 35° –Human FOV = 200 x 135° –Panoramic Mosaic = 360 x 180°
7
Recognizing Panoramas 1D Rotations () –Ordering matching images
8
Recognizing Panoramas 1D Rotations () –Ordering matching images
9
Recognizing Panoramas 1D Rotations () –Ordering matching images
10
Recognizing Panoramas 2D Rotations (, ) –Ordering matching images 1D Rotations () –Ordering matching images
11
Recognizing Panoramas 1D Rotations () –Ordering matching images 2D Rotations (, ) –Ordering matching images
12
Recognizing Panoramas 1D Rotations () –Ordering matching images 2D Rotations (, ) –Ordering matching images
13
Recognizing Panoramas
14
Overview Feature Matching Image Matching Bundle Adjustment Image Compositing Conclusions
15
Overview Feature Matching Image Matching Bundle Adjustment Image Compositing Conclusions
16
Overview Feature Matching –SIFT Features –Nearest Neighbor Matching Image Matching Bundle Adjustment Image Compositing Conclusions
17
SIFT Features SIFT features are… –Geometrically invariant to similarity transforms, some robustness to affine change –Photometrically invariant to affine changes in intensity
18
Overview Feature Matching –SIFT Features –Nearest Neighbor Matching Image Matching Bundle Adjustment Image Compositing Conclusions
19
Nearest Neighbor Matching Find k nearest neighbors for each feature –k number of overlapping images (we use k = 4) Use k-d tree –k-d tree recursively bi-partitions data at mean in the dimension of maximum variance –Approximate nearest neighbors found in O(nlogn)
20
Overview Feature Matching –SIFT Features –Nearest Neighbor Matching Image Matching Bundle Adjustment Image Compositing Conclusions
21
Overview Feature Matching Image Matching Bundle Adjustment Image Compositing Conclusions
22
Overview Feature Matching Image Matching Bundle Adjustment Image Compositing Conclusions
23
Overview Feature Matching Image Matching –Random Sample Consensus (RANSAC) for Homography –Probabilistic model for verification Bundle Adjustment Image Compositing Conclusions
24
Overview Feature Matching Image Matching –Random Sample Consensus (RANSAC) for Homography –Probabilistic model for verification Bundle Adjustment Image Compositing Conclusions
25
RANSAC for Homography
30
Overview Feature Matching Image Matching –Random Sample Consensus (RANSAC) for Homography –Probabilistic model for verification Bundle Adjustment Image Compositing Conclusions
31
Probabilistic model for verification
32
Finding the panoramas
35
Overview Feature Matching Image Matching –RANSAC for Homography –Probabilistic model for verification Bundle Adjustment Image Compositing Conclusions
36
Overview Feature Matching Image Matching Bundle Adjustment Image Compositing Conclusions
37
Overview Feature Matching Image Matching Bundle Adjustment Image Compositing Conclusions
38
Overview Feature Matching Image Matching Bundle Adjustment –Error function Image Compositing Conclusions
39
Overview Feature Matching Image Matching Bundle Adjustment –Error function Image Compositing Conclusions
40
Bundle Adjustment New images initialised with rotation, focal length of best matching image
41
Bundle Adjustment New images initialised with rotation, focal length of best matching image
42
Error function Sum of squared projection errors –n = #images –I(i) = set of image matches to image i –F(i, j) = set of feature matches between images i,j –r ij k = residual of k th feature match between images i,j Robust error function
43
Overview Feature Matching Image Matching Bundle Adjustment –Error function Image Compositing Conclusions
44
Overview Feature Matching Image Matching Bundle Adjustment Image Compositing Conclusions
46
Overview Feature Matching Image Matching Bundle Adjustment Image Compositing Conclusions
47
Blending
48
Gain compensation
49
How do we blend? Linear blending Multi-band blending
50
Multi-band Blending Burt & Adelson 1983 –Blend frequency bands over range
51
Low frequency ( > 2 pixels) High frequency ( < 2 pixels) 2-band Blending
52
3-band blending Band 1: high frequencies
53
3-band blending Band 2: mid-range frequencies
54
3-band blending Band 3: low frequencies
55
Panorama straightening Heuristic: people tend to shoot pictures in a certain way
56
Overview Feature Matching Image Matching Bundle Adjustment Image Compositing Conclusions
57
Overview Feature Matching Image Matching Bundle Adjustment Image Compositing Conclusions
58
Conclusion
59
Algorithm
60
AutoStitch program AutoStitch.net
61
Open questions Advanced camera modeling –radial distortion, camera motion, scene motion, vignetting, exposure, high dynamic range, flash … Full 3D case – recognizing 3D objects/scenes in unordered datasets
62
Credits Automatic Panoramic Image Stitching Using Invariant Features, 2007 –Matthew Brown and David G. Lowe (Uni. of British Columbia) Recognising Panoramas, 2003 –Matthew Brown and David G. Lowe (Uni. of British Columbia) –2003 –Thanks for the slides! Image Alignment and Stitching: A Tutorial, 2006 –Richard Szeliski (Microsoft)
63
Questions?
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.