CS 194: Computational Photography, Fall 2017

CS 194 Proj 6a: Image Stitching

Yunfan Song, CS194-acu



Overview

Part 1: Shooting and Digitizing

I used photos I have taken over the years. I split the photo of Shanghai for the purpose of the project.

Below are all the images I took:



Shanghai 1
Shanghai 2
Times Square
Cal 1
Cal 2

Part 2: Algorithm for Homographies

In order to rectify and warp images, I should recover the parameters of the transformation between each pair of images. In this case, the transformation is a homography: p’=Hp, where H is a 3x3 matrix with 8 degrees of freedom (lower right corner is a scaling factor and can be set to one). One way to recover the homography is through marking corresponding points in the two images. Essentially I need to solve 8 unknowns.

Part 3: Rectifying

When rectifying the small poster inside the picture, there seems to be a reflection




Unrectified Times Square
Rectified Times Square
Rectified Phantom of Opera Poster

Part 4: Blending

I morphed my images using alpha blending and I would add the left image to the left side of a large overall image. After some careful matrix multiplicaiton and addition, I then just summed the left and right images together.




Shanghai 1
Shanghai 2
Blend Shanghai skyline
Cal 1
Cal 2
My view walking on Piedmont Ave

Part 5: What I learned

I will never pay for photo editting Apps ever.