The system operating principle is saddled on the structured light triangulation. The structured light projection is a light form (plane, grid, etc…) and a known corner. By means of structured light the geometrical light information is exploiting in order to extract geometric information from the illuminated scene.
In Figure 1 we can see the effect introduced by a thickness object within a illuminatedscene by structured light consisting of a laser line projector. In case of bead research system, the laser plane intersection with the tube generates a circumference arc like the one visible in the cover page figure. In the bead proximity, the circumference is deformed due to the bead change thickness. The algorithms developed by Specialvideo allow to analyze the laser shape line and to accurately locate the bead position.
The solution advantage compared to traditional systems with contact or with distance sensors is to allow the bead identification even in ovalization or non-straightness tube presence.
One of the major characteristics of Specialvideo is to design and implement vision systems based on customer specific requirements. For this reason the system has been realized in various versions that differ primarily for the devices installation (cameras, laser, etc. ..) and for the use that is made of the vision system information. The system can be installed both inside and outside the tubes, in a fixed position or on the robot, and the coordinates detected can be used to command the robot position or to roll the tube in the desired position.
For this reason, performance and specifications may vary considerably from one version to another one. For example, the position measurements preaccuracy can go from the order of 0.1mm when the tube and the camera are in a fixed position, to 1mm when the tube or the camera are in motion and with large ra framed field.
To ensure high robot guidance systems measurements precision, Specialvideo has developed a procedure for the calibration of the the vision system with respect to the coordinates of the robot axes. This procedure allows in a automatic mode to: