Overview
This project was a fun foray into the world of manipulating images through Homographic matrices!
Technique
To create a panoramic image from multiple separate images, one must first define correspondence points between the image. For manually selected correspondence points, it is useful to pick distinct and easily recognizable objects present in both images. For the examples I chose, I selected edges and corners of image features, to ensure I selected the same points in both images.
The correspondence points are then run through a function to pick their homographic matrix, by compiling their points and running least regression on a few equations. After this matrix was calculated, the image to be warped was run through a warping function, resulting in an easily placable picture over the static image. This produced quite good results, as can be seen below
|
|
|
|
|
|
