Date of Award
2008
Degree Type
Thesis
Degree Name
Master of Engineering Science
Program
Mechanical and Materials Engineering
Supervisor
Ralph Buchal
Abstract
Laser scanners are widely used in quality control. Currently, the operation and programming of laser scanners is done manually. An efficient way of using this machine can save a considerable amount of time and money. This thesis develops several algorithms for the development of an off-line automated inspection planning system for laser scanners, focusing on inspection point visibility, clustering methods, and collision free path planning. A mathematical definition of the probe head motion and pose around the inspection point is presented. The visibility cones of given inspection points are calculated using graphics hardware rendering and image processing. Then the pixel images are geometric related to the probe head poses.
Three different clustering methods are developed with the objective to find the minimum number of probe head orientations required to scan all the inspection points. The methods are based on greedy and genetic algorithms. At last, a collision free path planning algorithm is presented. Equidistant strips in the direction of the laser axis are defined in order to scan the part.
The algorithms are implemented in a C++ program using OpenGL as a graphics library. The algorithms are tested on different areas of a work part defining the scanning paths required for its inspection.
Recommended Citation
Escoto, Abelardo, "INSPECTION POINT ACCESSIBILITY, CLUSTERING AND PATH PLANNING FOR LASER SCANNING" (2008). Digitized Theses. 4821.
https://ir.lib.uwo.ca/digitizedtheses/4821