Quite a few. Some I can think of are: 1) stereopsis (passive 3D point clouds and depth maps), 2) multi-sensor fusion for multi-spectral and hyper-spectral images and data cubes, 3) digital video stabilization, 4) panorama and mosaic stitching, 5) calibration. If you mean keypoint and feature vector analysis in general and not just simple registration, then there are no doubt more including a basis for segmenting scenes and recognizing objects in 3D for example.