When it comes to inspecting parts and packages, many engineers wish that machine vision cameras were more perceptive with regard to color. Until recently, the technology required to process large amounts of color image data was too slow for high-speed packaging lines, and not sensitive enough to detect coatings and metal finishes due to rainbow and shadowing effects. New technology, however, is moving color vision sensing closer to human perception
Vision inspection systems have come a long way in the relatively short time they’ve been around. Not long ago, they were limited to binary (black or white) or gray scale images and required complex illumination schemes to optimize contrast and differentiate objects. It wasn’t unusual for operators to continually adjust object position, orientation, lighting, and optics to achieve consistent and stable results. Even then, the results were limited in terms of resolution and contrast based on 256 shades of gray or black and white.
When color sensing was introduced, vision palettes were still limited to 256 hues and the color data had to be converted to grayscale to achieve faster processing speeds. Lacking accuracy, early color vision systems suffered from high false rejection and pass rates. Color inspection also lacked tools to help quality and maintenance people identify root causes of true faults and initiate timely corrective actions.
Near-human vision capability
The limitations of traditional vision systems have been overcome by using 24-bit “Truecolor” to mimic far more of the colors found in the real world and accurately capture complex images. A vision system color sensing engine called “Advanced Real Color Sensing” (ARCS) instantly processes up to 16.77 million colors in real time, using 256 shades of red, 256 shades of blue, and 256 shades of green. This translates into faster inspections without sacrificing quality, performance, or resolution.
Smart cameras improve results
As a way to reduce the impact of human error and limited operator skill, vision systems have offered cameras with “smart lighting” for about 10 years. Smart lighting simplifies setup and speeds changeover by allowing remotely selected repeatable settings. However, the built-in lighting sometimes does not illuminate the subject for optimal results and external lighting is required. Newer “intelligent cameras” have been introduced in the past year for expanded versatility and usability of the inspection system by plant personnel. These feature an all-in-one package, incorporating auto-iris, auto-focus, intelligent- LED lighting, and remote zoom control.
The auto-focus/auto-iris features reduce inspection errors caused by “accidental” changes to a manually adjusted lens. The left-hand cable image (above) shows the effect of an out-of-focus target, while the center image shows the same cables with an improper iris setting, both of which will cause inspection errors. Compare the right-hand image using the intelligent camera to the previous two images; the in-focus, properly illuminated object provides the operator with an image that nearly eliminates inspection errors.
Support tools help simplify
In addition to enhanced color sensing and smarter cameras, new support tools give operators the ability to address line problems — based on inspection results or input conditions — without involving engineering or shutting down the line. By using these tools, operators can make on-the-fly changes to measurement regions for work piece tolerances and switch between different measurement items based on the work piece. Additional tools help engineering and quality management get the most out of the real color vision system. These include testing of application viability using PC simulation software, automatic online justments to system settings based on feedback, and continuous automatic re-measurement applied to stored images for trend monitoring and process analysis.
Real world color sensing
Packagers of food and beverage products, pharmaceuticals, and cosmetics — who use color to attract attention on store shelves — have taken to analyzing their production processes in terms of Overall Equipment Effectiveness (OEE) to improve operations and profits. One tool that has made the transition easier is “real color” vision inspection, which gauges visual appeal (color consistency in printing) and information accuracy (date/lot code), and identifies package defects.
In automotive assembly and metal parts manufacturing, results from real color vision systems are used to substantiate and achieve certification for ISO/TS 16949, the global automotive industry quality system standard. Quick traceability in case of recalls and assurance of proper assembly help meet the goals of automotive tier customers. Particular inspection challenges that color vision systems solve include detecting thread sealants on bolts and confirming the formulation of rust/corrosion protective coatings on metal parts. Case hardened metal parts, such as auto seat tracks and brass fittings, can present a rainbow effect that confuses conventional vision sensors, causing a high rate of false signaling. Color vision is also used to sort out engines based on color and placement of distinguishing parts such as dipstick handles.
