reset password
Multi-Sensor Image Fusion (MSIF)
Phu Kieu, Keenan Knaur
Advisor(s): Elaine Kang

The purpose of the project is to develop an algorithm to fuse two images of the same scene from two space sensors and extract some meaningful information from that fusion. The algorithm is implemented using the OpenCV library which is an open source computer vision library.  The images used as input were data sets acquired from the GOES satellites operated by NOAA.  A rectification of the images was the first step and an equirectangular projection was used to accomplish this.  The images were then segmented to identify cloud pixels and then those pixels were grouped together to form cloud objects.  Those identified objects were then put through one of two matching algorithms to come up with a disparity map.  The disparity map was then used to calculate assumed altitude of the cloud objects over a certain range and the results are shown in a graph.

Resources
Project Documentation
Project Presentation (Slides)
Project Presentation (Video)