High performance optical triangulation ranging
Abstract
This thesis describes an investigation of the operation of optical triangulation range sensors and the effects which limit their accuracy. Sensors utilizing both coherent and incoherent light are studied. Theoretical and experimental results show that speckle, variable object surface reflectance, pixel size of the image detector, noise from the image detector, imaging magnification, and object surface structure all play a role in total sensor accuracy. Bounds on the degree to which each of the above factors affect accuracy are developed, and some current methods for limiting or reducing errors are mentioned. Improvements to the performance of optical triangulation sensors are made in two areas: sensor range accuracy when measuring a flat diffusely reflecting surface; and the ability to locate edges to within a fraction of an illuminating spot diameter. Sensor accuracy is improved with the development of image processing algorithms applied in a dual detector arrangement which normalize the effects of variable surface reflectance. Further, computer algorithms that relate detector image characteristics to edge location are developed. An experimental study determines which algorithms are best for edge detection, and what problems pose fundamental limitations to edge detection accuracy. Finally, the effect of surface roughness is examined to determine how this influences optical triangulation sensor performance.
Degree
Ph.D.
Advisors
Stevenson, Purdue University.
Subject Area
Mechanical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.