**Ming 1989 :** Most of beings have two eyes
to recognize their environment. But the basic principle of a vision system is
'one eye + motion'. In this paper, we first demonstrate a new proposition, that
is: **In machine vision, 3D scene perception needs necessarily two images, but
not two cameras. The basic structure of a vision system is one camera with its
motion (called 'dynamic vision')**.

We will then give one application of this proposition to the construction of a 3D polyhedral scene by a moving camera. This problem is decomposed into the following subproblems: (a) how to approximate a edge image by a set of 2D line segments; (b) how to match the set of 2D line segments within a sequence of images in order to obtain the dynamic parameters of 2D line segments; (c) how to recover the 3D line segments from the knowledge about the camera motion and the dynamic parameters of 2D line segments and (d) how to interpret the set of 3D line segments by a set of 3D polygons.