[go: up one dir, main page]
Scribd
Upload
82 views5 pages

Presentation - Image Stitching With RANSAC

Image stitching has applications in panoramic photography, aerial drone mapping, and medical imaging. It involves keypoint detection, finding image homographies using RANSAC, and stitching images together in sequence. The document demonstrates stitching two images, then multiple images sequentially, and compares the result to the commercial PTGui software. It also explains computing image transformations like translation, Euclidean, similarity, affine, and projective, and stitching images in a center sequence.

Uploaded by

Không Trân
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
82 views5 pages

Presentation - Image Stitching With RANSAC

Image stitching has applications in panoramic photography, aerial drone mapping, and medical imaging. It involves keypoint detection, finding image homographies using RANSAC, and stitching images together in sequence. The document demonstrates stitching two images, then multiple images sequentially, and compares the result to the commercial PTGui software. It also explains computing image transformations like translation, Euclidean, similarity, affine, and projective, and stitching images in a center sequence.

Uploaded by

Không Trân
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 5

Image stitching

with RANSAC
EAGLE EYE
Image stitching has applications in many places, such as
panoramic shooting in smartphone, high-precision map
synthesis for aerial photography of drones, and medical image
stitching. Image stitching with RANSAC has a wide range of
applications in these areas.

1.Keypoint detector: SIFT/SURF


2.Find Homography with RANSAC
3.Stitching images with Center Sequence

Step 1:Stitch two images


Step 2:Stitch images in sequence
Step 3:Compare the result with PTGui (Pro Stitching software)
Compute Transformations
𝑋 ′ =𝐻𝑋

[ ] [ ] [ ] [ ]
1 0 𝑥𝑡 co 𝑠 𝛉 − 𝑠𝑖𝑛 𝛉 𝑥𝑡 𝑎 0 𝑏1 𝑎 𝑏 𝑐
𝐻 𝑡𝑟𝑎𝑛𝑠𝑙𝑎𝑡𝑖𝑜𝑛 = 0 1 𝑦𝑡 𝐻 𝐸𝑢𝑐𝑙𝑖𝑑𝑒𝑎𝑛 = 𝑠𝑖𝑛 𝛉 𝑐𝑜𝑠 𝛉 𝑦𝑡 𝐻 𝑠𝑖𝑚𝑖𝑙𝑎𝑟𝑖𝑡𝑦 = 0 𝑎 𝑏2 𝐻 𝑎𝑓𝑓𝑖𝑛𝑒= 𝑑 𝑒 𝑓
0 0 1 0 0 1 0 0 1 0 0 1

{
𝑎 𝑥 𝑖 +𝑏 𝑦 𝑖 +𝑐

[ ] [ ][ ][ ]
′
𝑎 𝑏 𝑐 𝑥
′
𝑎 𝑏 𝑐 𝑥𝑖 𝑥 𝑖=  

{
𝑖 𝑔 𝑥𝑖 + h 𝑦 𝑖 +1 ′
𝐻 𝑝𝑟𝑜𝑗𝑒𝑐𝑡𝑖𝑣𝑒 = 𝑑 𝑒 𝑓 𝑦 = 𝑑
′ 𝑒 𝑓 𝑦𝑖 𝑥 𝑖 (𝑔 𝑥 𝑖 +h 𝑦 𝑖 +1)=𝑎 𝑥𝑖 +𝑏 𝑦 𝑖 +𝑐
𝑖
𝑔 h 1 𝑑 𝑥 𝑖 +𝑒 𝑦 𝑖 + 𝑓
𝑔 h 1 1 1 𝑦 ′𝑖 =   𝑦 ′𝑖 ( 𝑔 𝑥 𝑖 +h 𝑦 𝑖 +1)=𝑑 𝑥 𝑖 +𝑒 𝑦 𝑖 + 𝑓
𝑔 𝑥 𝑖 +h 𝑦 𝑖 +1
=
So it own 8 DOFs(Degree of freedom) and a 2D point has two DOFs
to (x,y) components
Stitching images with Center Sequence
4

2
1 3

Mid
zero one three four
two

1 2

Mid
zero two three
one
Stitching images with Center Sequence
Compare Our Panorama with PTGui

You might also like