This site will look much better in a browser that supports web standards, but it is accessible to any browser or Internet device.

Skip Navigation skip menu and banner
University of Wyoming

UW Technologies Available for Licensing

Technology Disclosure: 04-058 – Machine Vision And Real-Time Image Processing By Solid State Imitation Of A Biological System - Fly Eyes
 

Researchers have known for years that there are a host of limitations associated with the current state of the art in machine vision. Most of these shortcomings relate to the large amount of computations required with the current digital-based imaging systems, inadequacy of resolution, data transfer time from imaging sensor to host processor, and extensive processing time. Without improvement in one or more of these drawbacks, applications for machine vision will remain limited.

Current machine vision technology involves using a camera or sensor to capture an image (which may contain an object of interest), digitizing the image (or images, in the case of video), passing the digital data to a computer or processor, and performing algorithms and programs that parse the data and process them so that useful information can be extracted and used in specific applications. These algorithms and programs are computationally intensive – for example, data from sequential images have to be compared to one another in order to detect an object and find its position; further comparisons and calculations need to be done to determine whether the object is in motion and its direction and speed.

The amount of time required to continuously perform these tasks for every image means that important information might be missed. For example, during the time required to re-raster the viewing area, a fast-moving object may have already passed through without detection. Also, the sheer bulk of the computational power needed for current machine vision technology is problematic. Moreover, digital imaging dictates that the resolution of current systems is limited by the number of pixels – so often times, important details like whether there one or two fine objects in the viewing area cannot be answered. Taken together, these drawbacks limit the application of current machine vision technology.

In an entirely new innovation (see pending patent 2005/0279917), researchers at the University of Wyoming Department of Electrical and Computer Engineering have developed a machine vision system that is able to determine important visual information like the shape, position and velocity of an object in real-time. Taking cues from their studies of a biological system, our researchers have built a vision sensor that imitates what happens in the common housefly’s complex eye. By using lenses and photoreceptors arranged in a pattern reminiscent of a fly’s eye, as well as implementation of novel analog information processing algorithms (based on the electrical profiles of the fly’s ocular nerves), these researchers were able to immediately identify an object in the viewing area, determine its leading and trailing edges, its width, and its direction and speed - all in real time. This technology also has the potential to “see” and track multiple objects and is not dependent upon the background’s or object’s texture, contrast or color. Since it is analog, resolution is not sacrificed by the amount of pixels and each type of object will have a unique analog signal; thus, it is possible to catalog a multitude of different object signatures dependent upon the application of interest. This system is also scalable, so that field of view and depth can be customized to match the application. Also, the simple sensor developed can be expanded to one-dimensional array and two-dimensional array applications. See International Society of Optical Engineering Conference paper.

UW Researchers in Electrical and Computer Engineering Department were recently awarded $600K over a five year period to continue their investigation of fly inspired machine vision research (read more).

This University of Wyoming technology considerably improves upon the limitations of the current machine vision technology – computation intensity, resolution, and processing time. We feel that many more applications are now open to machine vision technology (military, manufacturing, field work, security, etc.). Below is a picture of an actual experiment in process showing the system detecting the movement of a thin wire across a black background in real time. 

If your company would like to learn more about this technology and how it can apply it in commercial situations, please contact Davona Douglass, director of technology transfer at the University of Wyoming. We would be pleased to enter a confidentiality agreement and share further details.