A basic problem in computer vision is to understand the structure of a real world scene given several images of it. The authors cover the geometric principles and their algebraic representation in terms of camera projection matrices, the fundamental matrix and the trifocal tensor.