The recovery of motion and depth from optical flow

The navigation of a robot in any environment requires the knowledge of the motion of the robot relative to the environment and of the three dimensional structure of the scene. This information can be obtained by using active sensors and/or by passive vision, provided by cameras mounted on the robot. In this paper it is shown how passive vision can be used to recover the time-to-collision and depth when the camera on the robot is translating. The proposed technique first computes the optical from a sequence of time varying images by using a modification of the algorithm recently proposed. The time-to-collision can be obtained by exploiting mathematical properties of the 2D motion field. Depth is obtained from the computed optical flow, using an equation already proposed by many authors.