Three-dimensional surface measurement by space encoding range imaging
Abstract
This article deals with a new method for obtaining the range image based on the triangulation method and surface description for 3-D object recognition. A Gray-coded pattern projector is used for space encoding and a CCD video camera is used for getting the image. In a set of binary images obtained from the Gray-coded-illuminated scene, the view-line orientation for triangulation is obtained from the address of the pixel and the projection angle is determined from the bit pattern of the corresponding pixel. Instead of scanning a slit light at 2 inch positions in turn, only n patterns are required for space encoding. As the Gray code has the property that the Hamming distance between the two adjacent numbers is always one, the ambiguity on the boundary of the projection pattern can be limited within 1 LSB width. Experimental results are presented. For 3-D object recognition, the surface description plays an important role as well as surface segmentation. This article describes two examples of surface representation: one using a quadratic surface fitting and the other using surface normal representation.
- Publication:
-
Journal of Robotic Systems
- Pub Date:
- 1985
- Bibcode:
- 1985JRoS....2...27S
- Keywords:
-
- Contour Sensors;
- Imaging Techniques;
- Pattern Recognition;
- Rangefinding;
- Robotics;
- Signal Encoding;
- Three Dimensional Bodies;
- Binary Codes;
- Image Processing;
- Projectors;
- Scene Analysis;
- Surface Properties;
- Triangulation;
- Engineering (General)