Panoramas are essentially multiple images fused together into a single image. The process of panorama creation from multiple images involves many steps; some are common to other computer vision tasks, such as the following:
- Extracting 2D features
- Matching pairs of images based on their features
- Transforming or warping images to a communal frame
- Using (blending) the seams between the images for the pleasing continuous effect of a larger image
Some of these basic operations are also commonplace in Structure-from-Motion (SfM), 3D reconstruction , visual odometry, and simultaneous localization and mapping (SLAM). We've already discussed some of these in Chapter 2, Explore Structure from Motion with the SfM Module and Chapter 7, Android Camera Calibration and AR Using the ArUco Module. The following is a rough image of the panorama creation...