In this chapter we discuss the spatial alignment of all spatial positions to a common coordinate system and forms the first stage in building a common representational format. To keep the discussion focused we limit ourselves to image registration defined as spatial alignment of two-dimensional images. The images themselves may belong to the same sensor type (monomodal registration) or different sensor types (multimodal registration). Image registration methods include: area matching (using mutual information), sparse parametric matching (using both handcrafted and deep-learning keypoints) and dense non-parametric matching (using both Lucas-Kanade and Horn-Schunck optical flow, SIFT flow and deep-learning flow). Finally we consider the problem of multiple image registration and bundle adjustment.