Automated vision systems are over 10 years old, but engineers are tapping recent technology advances find new applications. Often used with robots to find parts and orient tools, vision systems are also found where high-speed, accurate inspections are required along with data collection and networking.

A vision system has many components. One of the complexities of configuring a system is that many components can be combined into one box. Components include the object to be inspected, a camera with optics, lighting, frame grabber, image processor, display, software tools, and input/output devices. Sometimes, the frame grabber and the processor are one component. A camera with a built-in frame grabber has recently been introduced. The frame grabber, processor, display, software tools can be purchased as a unit. It is important in specifying and procurement that all components are covered.

The object to be inspected must be presented before the camera so that an image can be captured. Either the camera can be mounted on a robot arm as an aid to positioning, or the camera is mounted to a conveyor system with the part moving before it. Part placement relative to the camera is crucial. Although software tools have been devised to help the inspection tools find the part in the field of view, consistent placement from part to part assures consistent inspections and improves processing speed.

Lighting is usually the most important part of the entire system. Manufacturers provide a wide variety of fixture types and lamp qualities. Position of the light source often is the difference between a successful installation and great frustration. Backlighting helps define a shape in the image. Sometimes "washing" a part in direct light provides shadows that emphasize the area to be inspected. Colored light can help the system differentiate among various shades of color being inspected. A strobe light connected to a part present sensor allows inspection of rapidly moving parts.

Select a camera compatible with the frame grabber. Camera technology is rapidly improving. High-speed electronic shutters, built-in frame grabbers, and digital imaging are all available. Choosing the proper lenses will make image analysis much easier. All the physics learned in basic photography will be a life-saver here—like depth of field and field of view. Assure that the inspected characteristic is as clearly defined as possible in setting up the system to make the image analysis reliable and fast.

The frame grabber is essentially an analog-to-digital device. It converts the analog camera signal to a digital image. The image processor enhances the image and provides data for analysis and collection by the software tools. Some systems are proprietary, that is, the frame grabber, processor, and software are sold as a kit in a common enclosure. Often, a company provides a PC card frame grabber with software tools. The PC CPU (central processing unit) does the image analysis. A newer trend is to put a processor on the vision card. This frees the PC processor to handle other tasks like a spreadsheet importing data from the image analysis.

Software image analysis tools include the ability to count objects, compare shapes, and even measure a part. Optical character recognition (OCR) has become more reliable than in the past. The image of an alphanumeric character can be converted to its ASCII equivalent, which can be used in database or spreadsheet applications. Color recognition is constantly improving. Tools also provide a means for turning on an output for feedback to machine controllers or turning on an alarm. Many programs now read and interpret bar codes, including the new 2-D symbology.

Vision systems are becoming more powerful, yet easier to program. They are a useful tool to have in a complete machine control arsenal.