Link This |
Email this |
Blog This |
Comments (0)
Single-camera 3-D machine vision
March 28, 2006
For 3-D machine-vision applications, you can employ multiple- or single-camera approaches. As outlined by MVtec Software, the two-camera approach supports traditional variable stereo techniques and can serve in applications such as precision mold measurement. With this approach, a machine-vision system determines the depth of each single object point based on the disparity (distance from each of corresponding pixels) appearing in the transformed images from each camera.
However, often a single camera suffices to obtain 3-D images. The company reports that single-camera 3-D setups can yield fast results as well as sub-pixel accuracy. One method is based on analysis of features such as circular drill holes, which can appear as ellipses to an offset camera. Analysis of these ellipses enables the derivation of a 3-D image. Additional 3-D features can be derived by evaluating depth of focus and an object's reflectance.
The company's Website offers an example of a photometric stereo derivation of a 3-D image of soldering on a printed circuit board.
MVtec will demonstrate such techniques using its HALCON software at Hannovermesse (Hanover, April 24-28, 2006) and Automatica (Munich, May 16-19, 2006).
Posted by Rick Nelson on March 28, 2006 | Comments (0)
