Georgia Tech Research Horizons
TECHNOLOGICAL ACHIEVEMENTS Photonic Sensors
Computer-Aided Structural Design Artificial Vision
Virtual Therapy FalconView
Near-Field Antenna Measurements Millimeter Wave Radar
Near-Field Sampling Applied Chaos
Three-Dimensional Vision Solar Energy
Fused Silica Radomes Pulse Combustion
Monodisperse Aerosol Generation Digital Signal Processing
BASIC DISCOVERIES TABLE of CONTENTS


Technological Achievements at Georgia Tech

Artificial Vision

A machine vision device developed at Georgia Tech in the late 1980s is helping manufacturers improve productivity and quality by imparting greater accuracy and speed to their automated inspection capability.
Photo by Gary Meek
Manufacturers inspecting their products on the assembly line got increased speed and accuracy with Georgia Tech's development of an artificial machine vision system. Mechanical engineering professors Drs. Steve Dickerson and Kok-Meng Lee (pictured here) co-invented the device called SmartImage Sensor.

Marketed as the SmartImage Sensor, the device integrates optics, image-acquisition electronics and a microprocessor. The system can, for example, provide quality control inspection results and statistical process-control data and coordinate information for motion controllers.

The technology, invented by Drs. Steve Dickerson and Kok-Meng Lee of Georgia Tech's School of Mechanical Engineering, differs from traditional machine vision by eliminating TV signal-conversion electronics, fixed-frame rates and fixed field of view (image size and location).

SmartImage Sensor allows dynamic access of the charge-coupled device (CCD) with none of the horizontal image jitter associated with conventional frame-grabber-based systems. Horizontal jitter can result in image movement of more than 10 times the desired inspection accuracy. It also can lead to distorted geometric shapes and poor repeatability of performance, especially when performing measurement of components or other high-accuracy inspections.

The sensor also differs from traditional systems by providing a variable inspection rate of 40 to 70 parts a second; traditional systems have a fixed rate of 60 frames a second. SmartImage can be customized for specific manufacturing environments. Also, its embedded microprocessor can perform partial image acquisitions, increasing image-acquisition rates to more than 4,000 inspections a minute.

In 1991, Atlanta-based DVT Inc. was awarded exclusive licensing rights to the technology, and has continually upgraded the sensor's speed and image-resolution capabilities.

More than 4,000 SmartImage sensors are at work in nearly every manufacturing sector, including automotive, metal stamping, injection moulding, semiconductors, pharmaceuticals and electronics.

For more information, contact Dr. Steve Dickerson, School of Mechanical Engineering, Georgia Tech, Atlanta, GA 30332-0405. (Telephone: 404-894-3255) (E-mail: steve.dickerson@me.gatech.edu)


Contents    Research Horizons    GT Research News    GTRI    Georgia Tech

Send questions and comments regarding these pages to Webmaster@gtri.gatech.edu

Last updated: October 25, 1999