Advances in Color Vision Inspection
Color cameras are becoming more affordable and less complex.
Once upon a time, color vision technology was limited to large manufacturers with big budgets and lots of in-house technical expertise. But, recent advancements in cameras, lighting and software have made color vision more affordable, less complex and easier to use on assembly lines.
Vision technology is being integrated into a growing number of automation applications today because it enables new ways to solve complex production challenges.
Color provides critical information that can improve the reliability of many machine vision inspection applications. The ability to detect color can make it possible to inspect many products that cannot be inspected in grayscale. Typical examples include ensuring that a tan rather than a taupe automotive interior component is installed, and checking ink cartridges for leaks.
“Color vision allows users to monitor more conditions with the same device,” says Amanda Nelson, technical marketing manager for vision and vision lighting at Banner Engineering Corp. “For example, users can verify the color of an object in addition to its size and position. The question of ‘is the object the correct color?’ can now be answered, along with ‘is the part present?’ and ‘is the part placed in the correct location?’
“This is useful in applications where it is important to detect even slight changes in object color for quality control purposes,” explains Nelson. “In addition, color imagers can be used for inspections where color is the only difference between a pass or fail condition.”
“Color vision is important for assembly and inspection applications,” adds Yvon Bouchard, Asia Pacific technology director for vision solutions at Teledyne Imaging. “A color image represents the target accurately and quality control staff can relate to the image more easily than a grayscale image. Defect detection criteria also impose the use of color for some applications.
“The color response of materials can be enhanced, providing an increase in defect detection sensitivity,” Bouchard points out. “The use of colored light can further enhance images to highlight difficult-to-see defects.”
Vision systems that can detect color can verify that the correct components are present in an assembly. In fact, some inspection tasks, such as checking the color of wires in wiring harnesses, can only be done with color imaging technology. Other tasks, such as searching for objects moving along a high-speed conveyor, can also benefit from the additional information that color provides.
One of the main advantages of color processing is that it can detect and bring out subtle differences that grayscale imaging can’t. When a vision system captures grayscale images, it makes evaluation decisions based on a single value chosen among 256 shades of gray. When a vision system captures color images for evaluation, it has three times more data to work with.
According to Robb Robles, senior manager for vision product marketing at Cognex Corp., manufacturers typically use color machine vision technology to solve three main tasks:
Part sortation—color often is the only feature that differentiates some parts, such as caps or containers.
Color recognition and matching—humans often mismatch similar colors, while machine vision systems can reliably distinguish between them to monitor color for consistency.
Assembly verification and inspection—identifying that the right part has been used during assembly can be challenging when components are very small, or when identifying marks such as character strings or bar codes are not easily visible. In these cases, color offers an option for accurately determining if a part is assembled correctly.
“Today, with advancements in machine vision technology, imaging processing software and inspection tools, processing color images is becoming easier, faster and less expensive,” claims Robles.
But, not all machine vision applications require color processing.
“Color is not essential in most applications,” says Robles. “A grayscale camera can usually distinguish between colors based on different shades of gray.
“The vast majority of inspection applications are still best solved using traditional grayscale vision systems,” claims Robles. “However, there is growing interest in color imaging because the technology is much more affordable. Vision software also is easier to use and more accessible.”
“In some cases, you can get more reasonable image contrast and feature segmentation out of a grayscale image,” adds David Dechow, principal vision systems architect at Integro Technologies Corp. “One reason for this is because the algorithms that are used for machine vision don’t necessarily operate uniquely in a color space.
“For instance, processing a simple measurement task in color may not be of any benefit to manufacturers,” says Dechow. “But, if the task of the machine vision application is to differentiate, identify or segment by color, then a color system would make sense and would be the best way to go.”
An Affordable Alternative
Traditionally, color vision has been much more expensive than grayscale alternatives. But, that’s no longer the case.
“The biggest myth about color vision systems is that they’re too complex and too expensive,” claims Dechow. “On the contrary, standard color imaging today is not complex and is not expensive. However, advanced color imaging and precision color imaging systems that use multiple sensors are still complex.”
The main difference between standard grayscale and color cameras is a special array of color filters (Bayer filters) over the sensor pixels that provide the color images. This type of camera produces three images in a single acquisition: one each with red, green and blue content for the respective filtered pixels. However, while this feature reduces overall camera cost, it also reduces the camera’s spatial resolution.
“The standard Bayer filter color camera is now ubiquitous,” says Dechow. “Every vendor offers it alongside their traditional grayscale products.
“Color cameras cost virtually the same as grayscale or monochrome cameras,” explains Dechow. “There’s no real cost difference for color, because the technology needed to make the sensors is all at the chip level. Cameras that are able to provide color images use the same sensors and components as grayscale cameras.”
“Today, color vision systems use more complementary metal-oxide semiconductors (CMOS) instead of charge couple device (CCD) sensors because of the latter’s ability to operate at higher speeds,” says Felix Klebe, product manager for identification and verification products at Omron Automation Americas. “This lowers the cost and adds greater flexibility. That improvement has prompted more manufacturers to take a fresh look at color technology.
“We’re also seeing higher resolution cameras,” notes Klebe. “In the past, most color applications used 5-megapixel cameras. Today, more 12-megapixel cameras are available.
“The biggest myth associated with color vision involves cost,” Klebe points out. “You don’t have to invest a lot of money in hardware in order to benefit from this technology.”
“Color vision has become more affordable in recent years,” adds Nelson. “As the technology advances and demand increases, the cost of imagers, which are the framework of color imaging, is dropping. Faster processors and device memory enable machine vision systems to process the data faster, even with higher resolution imagers. Color vision has become a practical solution for everyday inspections.”
“Today, high-resolution sensors provide high-quality images in color, exceeding what was available in monochrome just a few years ago,” claims Bouchard. “The improvements in processing power, storage and cost make [affordable color technology] possible.”
As color vision inspection systems become more affordable, they can be deployed more widely on the factory floor.
“There’s no longer a big price difference between color and grayscale,” says Robles. “If an application really benefits from color, you can easily solve it.
“This has generated more interest from smaller manufacturers that traditionally avoided using machine vision technology,” Robles points out. “Lower-end vision systems are not nearly as complex as they were in the past. Small manufacturers can now configure a simple inspection application for less than $2,000.
“Vision software interfaces are increasingly easier to use,” claims Robles. “You don’t have to hire a programmer or a systems integrator. Small- and medium-sized manufacturers can solve their inspection problems quickly without a lot of vision experience. That’s now as true for color as it is for grayscale systems.
“The biggest myth about color imaging systems is that it involves layers of complexity,” says Robles. “The reality is that the technology is affordable and offers higher resolution and better performance than in the past.”
Lighting the Way
Lighting is important for all vision applications. But, it’s especially critical with color cameras.
“Lighting is the No. 1 challenge in color imaging applications,” says Dechow. “That’s because inconsistency in illumination changes the way that colors are represented in a scene.
“Color is only the wavelength of light that’s reflected off of an object,” explains Dechow. “So, if the lighting source is a white LED light that, by its nature, is a little more blue in its spectral curve, then shining that light on an object will result in an image that appears to be slightly bluer to the camera.
“Variation is also a challenge with lighting,” adds Dechow. “Factors in the environment, such as external ambient light, could change the color response of an object. Managing light can be very difficult in color vison applications.
“LED lights simply don’t have the capability of providing a broad spectrum and uniform response in spectral wavelengths over the entire visible light range,” says Dechow. “So, we get peaks and valleys in LED lights that can dramatically affect the color accuracy of some imaging applications.
“Ways to get around this include using balanced LEDs or old-school halogen lights, which deliver a much better color response,” Dechow points out. “But, there have recently been improvements in illumination systems designed specifically for color vision applications. They provide a more consistent and uniform spectral response.”
One recent illumination technology trend is called intelligent lighting.
“The objective here is to ideally exploit the reserves of illumination systems and also use them in a more efficient and economical way,” says Sophie Perrot, a sales engineer at Stemmer Imaging AG. “Intelligent lighting can be used to monitor illumination systems and is a step toward Industry 4.0. The technical properties and the dynamic application data of illumination are permanently available in the system.”
Omron recently unveiled new lighting technology that automates advanced defect detection and uneven color inspection as part of the latest iteration of its FH-series vision system. It can automatically detect small defects and subtle changes in color with great sensitivity. It can also simultaneously detect defects with different characteristics.
“One of the key technologies is advanced lighting technology that changes illumination angles, directions and colors depending on the specific characteristics of the object under inspection,” says Klebe. “Multidirection, multicolor (MDMC) light takes the automation of quality inspections to a whole new level.
“The vision system is designed to meet the growing demands for exceptional quality that are putting pressure on manufacturers to detect all scratches and other defects on products, no matter how small,” adds Klebe. “[In addition], today’s labor shortages are making it especially difficult to recruit enough visual inspectors to successfully complete all the necessary inspections without automating at least part of the process.
“While conventional vision systems have typically been unable to match the capability of the human eye for certain defect detection tasks, the FH series has surpassed this performance threshold with the MDMC light technology,” claims Klebe.
Other new products are making it easier for manufacturers to deploy color vision inspection technology.
Banner Engineering’s iVu Series vision sensors are now available in models capable of color inspections. They enable inspections based on color or variations in color, as well as by type, size, orientation or position. The compact, easy-to-use color vision sensors are available in models with an integrated or remote touchscreen.
Cognex’s In-Sight 9000 series now features the 9912, a 12-megapixel camera that can acquire detailed images for high-accuracy part location, measurement and inspection over large areas—even when mounted at long distances.
“This higher resolution enables you to inspect large parts or subassemblies, such as door panels or engine blocks,” says Robles.