Original Date: 11/03/1996
Revision Date: 01/18/2007
Information : Automated On-line Inspection
The Advanced Computation and Machine Vision Group at Oak Ridge National Laboratory (ORNL) is applying its experience and facilities in automated on-line inspection to uses other than weapons production. Its goal is to develop machine vision enabling technologies for high-speed applications with large data throughput. As a specialty laboratory, this Group is developing novel computational approaches and solutions, custom designing software and algorithms, and integrating it with off-the-shelf hardware. The result is a workable, field deployable prototype that has been specifically integrated to address solutions to difficult problems of national importance.
The developmental process flow used by ORNL involves the generation of specification requirements; reviewing available optical design and optical processing techniques; application of image processing, image analysis and human perception models; and integration of hardware and software to develop a system that effectively assesses the product, provides for production control, and enables the operator to respond immediately.
There are several projects currently being pursued at ORNL:
One system under development, sponsored through a Cooperative Research and Development Agreement (CRADA), is an on-line inspection system to assess workmanship defects such as chips, pits, inclusion, and dimensional measurement for ceramic substrates used in state-of-the-art electronic assemblies. The present process is accomplished manually. The semiconductor industry is also interested in the system for rapid automatic analysis and classification of defects on wafers.
Another developmental system will help the FBI modernize its fingerprinting process. The current process is a slow, manual operation that requires a significant amount of filing space. The system being developed at ORNL will digitize the fingerprint cards and allow users to identify suspects within seconds. The system scans in the present hard copy cards at a rate of 1,000 per hour (front and back).
Another system being developed for the Treasury Department and Postal Services will be used to inspect currency and postage stamps. These two products are also presently being inspected manually. The new process will utilize image signal processing, texture and color modeling, geometric processing, reasoning, and object modeling.
The machine vision program at ORNL is making significant technical advances in the area of industrial inspection for quality assessment and manufacturing control. It is developing enabling machine vision technologies for a broad range of applications of national importance. Through ORNL's innovative solutions to difficult inspection problems, it is providing production tools to the customer which enables them to be more competitive and productive.
For more information see the
Point of Contact for this survey.