This example shows how to create a panorama from a video sequence. The model calculates the motion vector between two adjacent video frames and uses it to find the portion of each frame that best matches the previous frame. Then it selects the matching portion and concatenates it with the previous frame. By repeating this process, it builds a panoramic image out of the video sequence.
The following figure shows the Panorama Creation model:
Motion Estimation Subsystem
This model computes the Sum of Absolute Differences (SAD) using Template Matching block to estimate the motion between consecutive video frames. Then it computes the motion vector of a particular block in the current frame with respect to the previous frame. The model uses this motion vector to align consecutive frames of the video to form a panoramic picture.
Panorama Creation Results
The model takes the video sequence in the Input window and creates a panorama, which it displays in the Panorama window. Note that this method of panoramic picture creation assumes there is no zooming or rotational variation in the video.
Available Example Versions
Some compressed multimedia files are only supported on Windows platforms. The examples that use these types of files can only be run on Windows.