A
laser range scanner is developed that acquires multiple-view range images of an
object and combines the images into a seamless surround image without having to
register the images. By manually sweeping a laser line over an object just like
painting over the object with a paintbrush, a range image of the object is
created. By appropriately adjusting the speed at which the laser is swept over
the object, dense points can be recorded in areas where high details are
present and sparse points can be obtained in areas where very little detail is
present. In the automatic mode, four lasers sweep over the object from four
sides in sequence while three cameras simultaneously record the laser stripes
to reconstruct the object. An implicit surface is fitted to the range points to
represent the scanned object by a smooth and shaded surface. Captured object
texture can be mapped to the created surface to construct a realistic model of
the object. The scanner scans an object from surround in 10 seconds and creates
a model of the object in 30 seconds. The scanner is portable and is self calibrating. Therefore, it can be set up and made
operational in a matter of minutes. The hardware for the scanner is also very
inexpensive.
An example scan obtained by the scanner is shown below. Fig. 1 shows three views of a manikin as seen by the three cameras during a scan. Fig. 2a shows the reconstructed range scan after implicit surface fitting, and Fig. 2b shows the surface after texture mapping. The scanner software provides tools to measure dimensions of the scanned object for quality control or other purposes. The tools shown in Fig. 3 can be used to electronically take measurements from a person's scanned body.

Fig. 1. Three views of a manikin during a scan as seen by the three cameras.

(a) (b)
Fig. 2. (a) An implicit surface representing the scanned manikin. (b) Reconstructed manikin after texture mapping.
Fig. 3. A measurement tool calculating different dimensions of the scanned manikin.
For more information about this laser range scanner, please contact Arthur Goshtasby, CSE Department, Wright State University.