Let's again use the camera that we calibrated in the first recipe of this chapter and take two pictures of a scene. We can match the feature points between these two views using the SIFT (short for Scale Invariant Feature Transform) detector and descriptor presented in Chapter 8, Detecting Interest Points, and Chapter 9, Describing and Matching Interest Points.
The fact that the calibration parameters of the camera are available allows us to work in world coordinates and, therefore, establish a physical constraint between the camera poses and the position of the corresponding points. Essentially, we introduce a new mathematical entity called the essential matrix, which is the calibrated version of the fundamental matrix that ...
