Active stereo

 Active stereo emerges as an alternative approach to the traditional use of two cameras. The word ``active'' here signifies that energy is projected into the environment. In an active stereo vision system, one of the cameras is replaced with a projector or a laser unit, which projects onto the object of interest a sheet of light at a time (or multiple sheets of light simultaneously) (Fig. 10). The idea is that once the perspective projection matrix of the camera and the equations of the planes containing the sheets of light relative to a global coordinate frame are computed from calibration, the triangulation for computing the 3-D coordinates of object points simply involves finding the intersection of a ray (from the camera) and a plane (from the sheet of light of the projector or laser).

  

Figure 10: Example of an active stereo vision system.

Calibration

In the calibration step, a calibration target with a number of known 3-D non-coplanar points relative to a global coordinate frame is set up in front of the camera and the projector. After adjusting the lighting and the focus of the camera and projector, the camera is first calibrated, the plane equation of each sheet of light projected by the projector is then computed.

• Camera calibration: this step involves recovering the $3 \times 4$ perspective projection matrix C. This matrix captures the intrinsic parameters of the camera, and the rotation and translation between the camera and the global coordinate frame defined by the calibration target.
• Projector calibration: for each sheet of light projected by the projector, compute the coefficients a, b, c, and d of the equation of the plane (relative to some global reference frame) that contains the sheet of light:

  a X + b Y + c Z + d = 0.

  At least 3 known and non-linear 3-D points lying on the plane are required.

Advantages and disadvantages

An active stereo vision system has the following advantages:

• it deals with scenes that do not contain sufficient features, such as edges or corners, which are associated with intensity discontinuities, for the stereo matching process
• the correspondence problem is totally absent. For each pixel in the image that is illuminated by a particular sheet of light from the projector or laser unit, the plane equation of the sheet of light would have been computed from the projector calibration procedure, simple triangulation is only required to compute the 3-D coordinates of the pixel.
• it gives a very dense depth (or range) map.
• it is applicable to many shape measurement and defect detection projects.

Its shortcomings are:

• the system must be pre-calibrated. Unlike stereo vision systems with two cameras, there exists no self-calibration technique for recovering the geometry between the camera and the projector (or laser) unit.
• for projector and camera pairs: well-controlled lighting is required, such systems are therefore restricted to working only in indoor environments.