3 October 1996 Neural-network-based noncontact measurement system
Author Affiliations +
Abstract
Now days, the most popular commercial non-contact measurement system uses the triangulation method. To obtain the accurate and quick results of measurement system, the calibration of the system is a must. In this paper, the ANN technology has been used along with laser scanning measurement systems. The image of the given object is measured with the help of a CCD camera and is fed as the input layer to the neural network, whereas the geometrical data of the object surface is the output layer. If the network is properly trained, it will provide the accurate surface fitting. The perceptron model has been adopted and trained by using an improved back-propagation algorithm. The results of the experiments have concluded the higher accuracy of the proposed measurement technique, as well as it is an easy and fast method. It has also eliminated the influence of lens aberration and other factors.
© (1996) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Limin Zhou, Limin Zhou, Abdul Waheed, Abdul Waheed, Bingheng Lu, Bingheng Lu, } "Neural-network-based noncontact measurement system", Proc. SPIE 2899, Automated Optical Inspection for Industry, (3 October 1996); doi: 10.1117/12.253079; https://doi.org/10.1117/12.253079
PROCEEDINGS
8 PAGES


SHARE
Back to Top