It is a side project for me to transform a normal(rectangular) image into a panoramic(circular) image. I found implementing through C++ is way faster than using Python. The concept is easy. We have to tell for each pixel on the target image, can where is the corresponding points on the source image.
Input image:
Output image:
- clone this directory
$ git clone https://github.com/alexunxus/rec2panoramica.git
- install opencv3
$ brew install opencv
If you cannot successfully install opencv, then checkout this page: https://stackoverflow.com/questions/34340578/installing-c-libraries-on-os-x
- open terminal and go to directory /rec2panoramica/
$ cd ~/My/Path/To/rec2panoramica/
- type make
$ make
If you have not yet install GNU make, then
$ brew install make
- put your image into this directory and type
$ ./myImg yourjpgname
- your image will be stored in the same directory with prefix "pan"