Method and system for online monitoring and control of beverage can color decoration specifications

Abstract

The present invention provides a system for automated machine vision inspection environments. The system includes an inspection camera and a spectrophotometer or spectrometer, both of which are implemented for in-line use to detect the absolute color of printed portions of an inspected article. The spectrophotometer or spectrometer is aimed at a fixed light spot within the field of view of one of the inspection system's digital cameras, with the inspection system having a priori knowledge of the spectrophotometer's or spectrometer's exact aiming position. Images captured by the camera are used to determine whether the light spot to be measured on a decorative pattern has actually been measured using the most recent snapshot of spectrophotometric data. When the vision system determines that the spectrophotometer or spectrometer was truly aimed at the correct area when it captured its inspection data, the vision system instructs the system to accept the color measurement and records the relevant data and information accordingly. If the correct light spot was not measured, the data can be simply discarded or retained for other uses.

Description

[0001] This application is based upon and claims priority from U.S. Provisional Application No. 62 / 632,955, filed on February 20, 2018, which is incorporated herein by reference in its entirety. Technical Field

[0002] The presently described embodiments relate to color decoration of beverage cans or other generally cylindrical objects when they are manufactured at high speeds. This can include a wide range of consumer product packaging, such as caulking tubes, frozen juice containers, pudding cups, and beverage containers. Background Art

[0003] In the 2-piece beverage can industry, aluminum and steel cans are printed or "decorated" around their peripheries on high-speed rotating machines (generally referred to as decorators or decorator machines or systems). Although the basic design configuration of decorator machines has not changed substantially in more than 40 years, they have undergone continuous evolutionary fine-tuning to allow the machines to operate at higher and higher speeds. Some of the highest speed decorators can operate well at over 2,400 cans per minute. The latest trend is that instead of simply increasing the speed of a single can decorating lane, the line will be split and handled by two slightly slower decorators.

[0004] Regardless, the process proceeds too quickly to be observed by the human eye, which has any depth of perception. The human eye is incapable of performing a "stop-motion" function. Humans typically attempt to synchronize their vision with the movement by moving their heads and eyes, but such synchronization attempts are ineffective at higher speeds, and certainly not at 2,400 articles per minute. Even at the speeds anticipated by the decorator system and accompanying machine vision technology (e.g., over 2,400 articles per minute, as described above), with the aid of strobe lighting, it would be impossible for a human viewer to derive the necessary information to determine whether the printing process is operating correctly and within process control limits. Even if a human could see the cans in this environment with the aid of strobe lighting, it would still be impossible to adequately manually inspect the can decoration at the speeds typical of modern decorators. Human vision, while a remarkable device, cannot achieve the absolutely definitive color specifications that can only be verified with calibrated instrumentation. Even if an operator could somehow view the cans at sufficient speed, they would not be able to see 360° around the can from a vantage point to verify all aspects of the printing.

[0005] In recent years, multi-camera high-speed vision inspection systems have been used to monitor the quality of can decoration. This high-speed vision inspection technology is described in U.S. Patent Application No. 13 / 277,157, filed on October 19, 2011, and the U.S. Patent Application No. 13 / 277,157 is incorporated herein by reference in its entirety.

[0006] In order to perform a comprehensive inspection of the can 360° around its perimeter, the periphery of the can image is typically "unrolled" in some form. There are other ways to view its perimeter, but this is the most practical. Essentially, a plan view that appears to be rolling out the can allows for a more comprehensive view. This can generally be done in at least two ways. A line scan camera can be used to capture a series of (1xN) long single pixel rows or long, very narrow images (e.g., vertical pixel rows or narrow line images) that, when reassembled in a computer or vision processor electronics, will approximate an unrolled image of the exterior of the can's decorative surface. This unrolled image can be considered an image "pseudo-blanket" due to its imaging similarity to an actual printing blanket that applies ink to the perimeter of the can.

[0007] A true image of the actual printing blanket will produce an inaccurate image compared to the decorated can image. Due to the compressibility of the polymer from which the printing blanket is made, the actual dimensions will be slightly different compared to the printed can. While the ink color is visible on the surface of the printing blanket itself, it is not a true representation of the color when the ink is actually applied to the can surface. This is because the printing blanket is a polymer with a relatively matte finish and has its own color, rather than a highly reflective surface like the metal from which the can is made.

[0008] Another way to provide an expanded image is to take multiple, more generally proportioned, "zone scan" images—for example, a 3:4, 16:9, or even 1:1 aspect ratio of the outer perimeter, and then perform a correction and stitch these zone images together to form a "pseudo printing blanket" or inspection image. Many different algorithmic techniques exist for opening, expanding, and stitching images together, but a comprehensive review and explanation of those techniques is beyond the scope of this invention. Once the image is reassembled or at least pulled into a computer or image processor in a form that can be used for inspection, the process of analyzing the image to determine print quality must begin. It may be desirable to inspect many different aspects of the can's decorative surface. An inspection system may be employed to verify that any of dozens of different printing features have been correctly applied to the can.

[0009] The actual machine vision algorithm technology used to interrogate the image of the can to determine the presence of defects is up to the software designer and the electronic processing architecture. The details of the actual visual inspection in the vision processor or computer are also beyond the scope of this invention.

[0010] The inspection system can be used to inspect for, for example, smudges, misalignments, oil spots, blanket tears, swirls, wrinkles, contamination, incomplete ink transfer, insufficient ink, missing letters / characters / features, and various other important aspects related to correct color. As will be described in more detail below, the subject matter of the present invention is primarily directed to a novel way of verifying correct color.

[0011] Absolute color is crucial for a variety of reasons. Brand owners often rely on precise color reproduction to remain consistent throughout their advertising campaigns, so that simply seeing that color brings their product to mind. Some popular soda brands employ this approach and therefore expect color reproduction on their products to be not only accurate but also highly consistent and transportable across all production lines worldwide. By implementing automated inspection technology on print decorators, manufacturers have realized how variable the final decorated color can be. While many once believed that color in a decorator was relatively stable over relatively long periods of time, monitoring and inspecting each can has revealed that, in addition to well-known long-term trends, color can also vary over relatively short periods of time. Color can vary, either within or outside of specifications, for a variety of reasons, often depending on which machine sections are actually responsible for decorating each individual can.

[0012] Ink manufacturers will typically publicly explain that a lengthy list of over 40 different reasons, either individually or in combination, can lead to significant variations in the actual measured color on a decorated can. The color can be affected by over three dozen independent variables, ranging from the condition of the decorator to the condition of the metal substrate, to the details of the decorating process, to the ink itself, to the condition of the varnish, to the human factors of setup, measurement, and observation.

[0013] State-of-the-art visual inspection techniques are typically implemented using digital color cameras. While the cameras aren't used to capture continuous-frame video information, they serve as a digital, electronic means of acquiring high-speed (short-exposure-time) individual images of the product. While extremely intense continuous illumination techniques can be used, enabling the camera to utilize a high-speed electronic shutter, this is typically a more expensive and impractical solution. In most cases, high-speed images can be captured by utilizing strobe lighting to provide intense pulses of light for very short durations to "stop the action" and thereby capture a clear, properly timed image with minimal blur into the camera. The illumination on-time is typically less than 150 microseconds but can range widely depending on the movement and throughput speed of the decoration inspection system. It is well known in the art to use engineered lighting to produce images with the most desirable characteristics for a specific inspection application. In most cases, the angle, intensity, relative uniformity, and color content of the illumination are critical to high-quality machine vision inspection. When inspecting color decorations on cans or inspecting the entire 360° circumference of a cylindrical container, a high degree of engineering design is required to avoid hot spots or specular reflections and to achieve consistent image quality across all areas to be inspected. The inks used to decorate many modern cans are translucent, so the highly specular nature of the underlying aluminum or steel surface becomes a significant consideration in lighting design. Lighting designers face a significant dilemma. Brightfield specular lighting is challenging to achieve because the mirrored cans will often cause the image of the actual light source to be imaged, and therefore must be near-perfect. Furthermore, the typically inherently curved outer perimeter of a beverage or food can acts as a reflector for the colored light from the light source near the camera's focal plane. The result is that neighboring cans are "imaged" in the mirrored surface of the can, which is currently the subject of inspection. Designing a light source that can be continuous around a 360° perimeter is impractical because the stream of cans must move through the illumination field. At these extreme speeds, consistently moving each can in and out of a perfectly rotationally uniform illumination field is impractical and expensive. Furthermore, due to the relatively high illumination intensity, brightfield illumination washes out most color and printed pattern information. Any deviations from a perfect cylinder on the can's outer perimeter (for example, such as dents or structures) are magnified and significantly distort many aspects of the image, including color information. It's a bit like illuminating a curved, colored mirror. With off-axis lighting from above or below the can, much of the illumination incident on the can's curved, mirror-like surface is simply reflected away from the camera. In reality, only a small fraction of the light is reflected by the pigmented particles in the thin body of actual ink to be used by the camera to create a color-reproduced image.

[0014] Today's color cameras, which are commonly used in beverage can inspection systems, have significant inherent limitations. These commonly used cameras are fundamentally called tristimulus devices. There are several different types of color digital cameras, but they generally use two, three, or more color filters to determine the color information used for imaging. While custom filters are possible, almost all custom filters use standard red, green, and blue filters, thus dividing all colors of visible light entering the camera into one of those three color buckets. The filters are selected to attempt to approximate how the human eye perceives color. With good lighting, setup, adjustment, and calibration, these filters can do a very reasonable job of color imaging. If implemented correctly, an inspection system that uses its tristimulus color data can do a very reliable job of monitoring color changes, variations, or process shifts.

[0015] For example, a three-chip color camera utilizes three imaging chips, one dedicated to each of the red, green, and blue color bands. The imaging chips are aligned with mirrors, optics, or prisms so that corresponding image information falls on corresponding R, G, or B image planes in corresponding locations. The three imaging array chips must be precisely aligned or calibrated with each other pixel by pixel, which is one reason this type of camera is much more expensive. In this way, a true numerical intensity or signal strength value is generated for each pixel in each of the three color image planes. In total, there are only three numerical values ​​representing the color information for each pixel.

[0016] Single-chip color digital cameras use a different approach. There are different technologies for achieving this, but the most common technique incorporates a Bayer filter. This type of single-chip camera arrangement divides the entire pixel matrix array on the imaging chip into groups of four pixels. A color filter is overlaid on each of those pixels (a so-called Bayer filter), so that there are two green-filtered pixels, one red-filtered pixel, and one blue-filtered pixel. This forms a 2x2 array, with two green pixels positioned diagonally opposite each other with their corners touching, and red and blue filling the other two adjacent positions. This means that only one of the four pixels is sensitive to red band information, while two of the four pixels see green information, and only one of the four pixels sees blue information. Technically, and especially in detail, it is obvious that there is only incomplete color information available for each pixel, or even each 2x2 pixel group. This results in various data gaps between pixels and can cause image aliasing and moiré patterns. A single-chip arrangement results in reduced resolution, poorer color fidelity, and more limited dynamic range, but is a much cheaper way to manufacture a color camera.

[0017] Regardless of the type of color camera used, only three pieces of digital information are generated to attempt to represent the exact color from each pixel or group of pixels. Sometimes a calibration matrix or tuning matrix is ​​used in the process of calculating the 1x3 data array for each pixel, but the data is originally derived from and ultimately becomes a 1x3 array of values ​​that represents the relative intensities of red, green, and blue color information.

[0018] The visible color spectrum is defined by a range from approximately 400 nanometers to approximately 750 nanometers. If this 350 nanometer range is divided into three segments for each of the tristimulus filters, each range is still greater than 115 nanometers. Color filters are typically designed not to have perfectly sharp cutoff points, so for practical manufacturing, they must typically overlap or have valleys or gaps between filters. Using one digital data segment to represent the entire variable complement of color data within the 115 nanometer range does not allow for a truly accurate representation of the color signature. Even within the range of one of the color filters, the color signature curve can still be very large. If, as is the case with most color signatures, the color signature curve overlaps with two or more of the color filter ranges, then the color signature curve is not an accurate representation of the color signature curve. The more complex the color signature curve, the less likely it is to be correctly represented by a tristimulus-based sensor system.

[0019] However, it is well known in the color community that tristimulus color cameras are no match for spectrophotometers or spectrometers when it comes to absolute color measurement and comparison to color specifications. For example, a spectrophotometer or spectrometer can be, and often is, calibrated so that it is traceable to a NIST color standard. Unlike the cameras described above, a spectrophotometer or spectrophotometer divides the 350-nanometer visible light range into more than three data segments. Spectrometers and spectrophotometers have had many different resolutions and specifications over the decades that they have been constructed. For the purposes of this disclosure (including the appendix), the terms spectrometer and spectrophotometer will be considered interchangeable, even though the latter typically includes an engineered illumination source, while the former typically uses available light from a variety of available sources. Depending on the intended use and price point of the instrument, the visible spectrum can be divided into 16 to several thousand segments. Even inexpensive spectrophotometers or spectrometers currently typically have 1,024 data points to describe the color signature curve of the incoming visible light. Obviously, with such fine data resolution, the color signature curve can more accurately represent the exact color and compare the exact color to an accurate standard or specification. A well-designed spectrophotometer or spectrometer, if properly calibrated and used, can be traceable to NIST color standards and can provide transportable color measurements that can be used to ensure worldwide color standards for manufacturers around the world.

[0020] Furthermore, so-called “single-chip” spectrometers are becoming increasingly available. Single-chip spectrometers effectively shrink a spectrometer to a tiny fraction of the size of a conventional spectrometer. This can be accomplished in a variety of ways, but one approach is to use a two-dimensional imaging chip with a color mask over the photosensitive area of ​​the chip that incorporates a stepped matrix with slightly different color filters for each pixel. An imaging lens focuses light from a common point to be inspected onto the entire (or portion of) the imaging chip so that only light from the corresponding filtered color reaches its corresponding pixel. Typically, some type of averaging filter or integrating sphere is employed in the light path so that the light diffracted onto the sensor all has a spectral content similar to the average value of the measurement “spot.” The color signature of the spot of interest is then expressed as an array of values ​​for the signal output from the entire set of pixels corresponding to each color filter. Thus, if the set of color filters represents the entire range of the spectrum to be measured, then the signal signature represents the measured spot.

[0021] Because imaging chips can be made very small using current technology, the size of the entire spectrometer, including the imaging lens and light integrating sphere, can be less than a few cubic millimeters. The size can vary depending on the resolution and light-gathering capability. The greater the light-gathering capability, the faster the spectrometer can integrate enough light to obtain a good reading at a given illumination intensity level. It is important to use a spectrophotometer configuration that has sufficient light-gathering capability, sufficient resolution, and sufficient speed to operate appropriately in the desired application. In general, it is possible to trade off resolution for additional speed. For example, at a given sensitivity level, if twice as many pixels are used to collect light in each spectral range, the device should have enough light to meet the minimum signal-to-noise ratio in half the time. This concept can be extended to allow originally slower spectrometers to operate at higher speeds while still providing sufficient spectral data resolution for a given application.

[0022] A spectrophotometer or spectrometer must use correct and consistent lighting to obtain viable color measurements that are traceable to NIST, corporate, or other color standards. While it's beyond the scope of this invention to teach the important lighting aspects that must be part of a good design, they will be briefly reviewed here. This information is well known in the art and can be referenced in depth elsewhere. Generally, spectrophotometers or spectrometers should operate without specular illumination. That is, illumination that could produce direct specular reflections from the surface being measured should be completely avoided. The best lighting scheme is one that employs soft, uniform illumination. For in-line beverage can or cylindrical object inspection, truly uniform illumination is often impossible, but the next best thing is the right kind of oblique (non-specular) illumination. Another standard commonly used in spectrometry is soft, approximately 45° illumination. This assumes that the primary component of the illumination will be incident from above or below the item at an angle of approximately 45°. When implementing this lighting scheme, designers should be careful to avoid surfaces within the illumination field that would produce hotspot specular reflections, which could ultimately negatively impact the accuracy and consistency of color readings. While it is nearly impossible to keep the lighting completely pure with respect to these issues, ultimately better results in the inspection / measurement system will make up for any effort.

[0023] In summary, imaging the perimeter of a can decoration in order to obtain truly accurate, transportable, and repeatable measurements of the absolute color in a known area of ​​the can's decorative label or pattern is quite challenging. It is manually impossible to accomplish this task. Digital color cameras are insufficient to perform this task with the highest levels of accuracy and repeatability. Using a spectrophotometer or spectrometer at high speed is often impossible when the exact aiming point that must be measured is not necessarily within the instrument's field of view due to the random orientation of the beverage can decoration. Summary of the Invention

[0024] In one aspect of the presently described embodiments, a system for monitoring and controlling color decoration on a rotating, non-oriented cylindrical object comprises: a spectrophotometer or spectrometer; and an image-based machine vision inspection system including at least one camera, an illumination system, a microprocessor or computer-based processor, and timing / triggering / control electronics, the image-based monitoring inspection system being configured to capture an image of the perimeter of a color-printed cylindrical object in a field of view and feed image data to the processor; wherein the spectrophotometer or spectrometer is configured to obtain a light point color reading within the field of view where the image is being captured by a camera of the image processing monitoring system and feed the light point color data to the processor; and wherein the processor is configured to: utilize a priori knowledge of the coordinates of the location where the light point color spectrophotometer or spectrometer reading is located within the field of view of the camera to use at least one image logic algorithm to determine the exact position orientation of the color decoration on the cylindrical object imaged in the field of view, use logic to determine whether the color light point data of the spectrophotometer or spectrometer is from the desired location within the image, use logic or rules to determine whether the light point color data can be used to compare with the actual orientation of the sampled captured image data, and output high precision color data based on the image data, the light point color data, and a determination of at least one of monitoring the manufacturing process and correcting the manufacturing process.

[0025] In another aspect of the presently described embodiments, the cylindrical object is a colorfully decorated container.

[0026] In another aspect of the presently described embodiments, the cylindrical objects are a stream of non-rotating oriented, colored decorated 2-piece beverage or food cans in the manufacturing process.

[0027] In another aspect of the presently described embodiments, the a priori knowledge of the coordinates of the light point color readings of a spectrophotometer or spectrometer is determined by equipping a spectrophotometer or spectrometer lens arrangement with a laser that makes a recognizable mark on the camera's image.

[0028] In another aspect of the presently described embodiments, the processor has logic to locate the identifiable marker and record the corresponding coordinates for future verification when the laser is not turned on.

[0029] In another aspect of the presently described embodiments, the processor uses the light point color data from the spectrophotometer or spectrometer to perform at least one of verifying camera color information, recalibrating the camera, and amplifying camera color output.

[0030] In another aspect of the presently described embodiment, the processor uses additional logic algorithms to check calibration on all system cameras by statistically sampling and comparing recent color readings obtained by the spectrophotometer or spectrometer from non-simultaneous images formed by the cameras, the non-simultaneous images including the photosite region of interest of the spectrophotometer or spectrometer.

[0031] In another aspect of the presently described embodiments, the processor sends a signal to move the spectrophotometer or spectrometer position aiming point parallel to the central rotational axis of the cylindrical object so that another vertical band is monitored as the cylindrical object reaches the inspection position at various rotational positions.

[0032] In another aspect of the presently described embodiments, an array of more than one spectrophotometer or spectrometer is connected to the processor, each spectrophotometer or spectrometer in the array focused on a different inspection band on the cylindrical object.

[0033] In another aspect of the presently described embodiment, an array of spectrophotometers or spectrometers is provided, each spectrophotometer or spectrometer in the array focused on a different inspection band around the circumference of the cylindrical object, so that multiple bands of interest can be monitored without moving the spectrophotometer or spectrometer from a fixed position.

[0034] In another aspect of the presently described embodiments, the array of spectrophotometers or spectrometers may be moved as a unified unit so as to cover a desired area of ​​the color decoration but keeping the spatial relationship between the spectrophotometers or spectrometers fixed.

[0035] In another aspect of the presently described embodiments, the spectrophotometer or spectrometer position can be one of: manually, semi-automatically, and automatically moved to a position so that its focused light point is in the field of view of a different camera.

[0036] In another aspect of the presently described embodiments, the high precision color data is displayed so that it can be used to manually correct the decoration manufacturing process.

[0037] In another aspect of the presently described embodiments, the high-precision color data is output to effectuate automatic correction of a decoration manufacturing process.

[0038] In another aspect of the presently described embodiments, the output is connected to a control system that can directly drive servo adjustments to at least one of a digital inking head, additional print heads, print wheel adjustment, and individual spindle adjustment arrangements.

[0039] In another aspect of the presently described embodiment, the high-precision color data based on the spectrophotometer or spectrometer is processed and organized so that the system can be initialized to output the data in the most useful form of the data, which may be one of CIE XYZ, CIE Lab, CMYK, and CIERGB.

[0040] In another aspect of the presently described embodiments, the high precision color information is processed and output to determine when there is a statistically significant difference between color quality standards that have been input into the system and recent actual production color readings.

[0041] In another aspect of the presently described embodiment, a method for monitoring and controlling color decoration on a cylindrical object in a system having a spectrophotometer or spectrometer and an image-based machine vision inspection system, wherein the image-based machine vision inspection system includes at least one camera, a lighting system, a microprocessor or a computer-based processor, and timing / triggering / control electronics, comprises: capturing an image of the perimeter of a color-printed cylindrical object in a field of view; feeding the image data to the processor; obtaining a light point color spectrophotometer or spectrometer reading within the field of view where a camera of an image processing monitoring system is capturing the image; feeding the light point color data to the processor; determining the exact position orientation of the color decoration on the cylindrical object imaged in the field of view, determining whether the color light point data is from the desired position of the decorated label within the image, determining whether the light point color data can be used for comparison with the image data, and outputting high-precision color data based on the image data, the light point color data, and a determination of at least one of monitoring the manufacturing process and correcting the decoration process.

[0042] In another aspect of the presently described embodiment, a system for color inspection of generally cylindrical containers that are color-decorated around their periphery includes: a machine vision inspection system having at least one imaging camera configured to capture an image of at least a section of the decorated periphery of the container; a spectrophotometer or spectrometer configured to collect precise color information from a desired area of ​​interest of the decorated periphery where the camera has captured an image; and at least one processor configured to compare the color authenticity of the image with the color information collected by the spectrophotometer or spectrometer, and use the comparison to check the color accuracy of the camera if the desired section from which the spectrophotometer or spectrometer collected color information is in a desired location on the decorated label.

[0043] In another aspect of the presently described embodiment, a method for color inspection of a generally cylindrical container that is color-decorated around its periphery includes: capturing a color image of at least a section of the decorated periphery of the container; collecting accurate color information using a spectrophotometer or spectrometer from a desired section of the decorated periphery where the camera has captured the image; comparing the color authenticity of the image with the color information collected by the spectrophotometer or spectrometer; and using the comparison to check the color accuracy of the camera if the desired section from which the spectrophotometer or spectrometer collected color information is in a desired position.

[0044] In another aspect of the presently described embodiment, the method further includes at least one of: using the color comparison information to perform one of correcting calibration of the camera and modifying the camera color output; and using high-resolution spectrophotometer or spectrometer color information to monitor and correct at least one of the color decoration process.

[0045] In another aspect of the presently described embodiments, color settings of the camera are recalibrated as a result of the comparison with the readings from the spectrophotometer or spectrometer.

[0046] In another aspect of the presently described embodiments, other cameras in a multi-camera cylindrical container inspection system are also recalibrated, wherein the computer / processor statistically reviews the color readings of cameras in the same decoration location to determine possible recalibration settings. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a diagram of an example system according to the presently described embodiments;

[0048] Figure 2 is a diagram of an example system according to the presently described embodiments;

[0049] Figure 3 is a diagram of an example system according to the presently described embodiments;

[0050] Figure 4 is a diagram of an example system according to the presently described embodiments;

[0051] Figure 5A is a diagram of an example system according to the presently described embodiments;

[0052] Figure 5B is a diagram of an example system according to the presently described embodiments;

[0053] Figure 6 is a diagram of an example system according to the presently described embodiments;

[0054] Figure 7A is a flowchart illustrating an example method according to the presently described embodiments; and

[0055] Figure 7B yes Figure 7A Continuation of the flowchart. DETAILED DESCRIPTION

[0056] Digital camera-based inspection systems, as discussed above, do not have the ability to measure color with the highest levels of repeatability and transportability. However, with the proper implementation of various orientation algorithms, they do have the ability to determine the exact orientation and position of each individual can label as it is being inspected. A spectrophotometer or spectrometer, on the other hand, will generally not be usable for pattern conditions that change at high speeds because it does not know anything about the orientation of the can's label decoration, i.e., it has no spatial information. It simply integrates visible light reflected from its target field of view and measures the exact amount of color information entering the instrument at each of its wavelength bins. The presently described embodiments relate to a combined "smart" system that includes both a camera-based inspection system and a spectrophotometer or spectrometer. This combined "smart" system will most typically incorporate at least one color imaging camera, but it is possible to implement this concept using a non-color grayscale camera. According to the presently described embodiments, the combined system is intelligent or "smart" because it is able to determine exactly where a spectrophotometer or spectrometer is being aimed at or viewed relative to a printed pattern on a cylindrical object, such as a colored decorated container (e.g., a beverage can or a 2-piece beverage or food can), and is able to correlate that location with the data or color value that should be measured at that location.

[0057] As an example, if seven different basic colors are being used to decorate or print this beverage can, it is desirable to measure each of those colors at one or more locations on the peripheral surface of the can. Spectrophotometers or spectrometers are typically capable of viewing a limited spot diameter and a fixed position. Enlarging or reducing the spot size is possible, but changing the spot size is not a typical approach to utilizing a spectrophotometer or spectrometer. The spectrophotometer or spectrometer is typically calibrated to a specific spot size, measuring at a specific angle, at a specified distance, and utilizing calibrated illumination of known characteristics or color temperature mixtures. All of these parameters are crucial for the design of a spectrophotometer or spectrometer to ensure consistent measurements. If it is desired to have color measurements traceable to NIST standards, then more careful calibration to consistent industry standards is necessary.

[0058] When using a spectrophotometer or spectrometer in a can decoration color measurement application, one of the problems solved by the presently described embodiments is as follows. If the spectrophotometer or spectrometer is set up to see a spot size of 1 / 4 inch in diameter, then it must be aimed at an area of ​​the can decoration that has only the specific color in question within the entire 1 / 4 inch diameter spot or field of view that the spectrophotometer or spectrometer is aimed at. Good practice is to have a margin around the actual field of view to ensure that any slight misaiming still ensures that the measurement is not contaminated by surrounding colors. For consistency, good practice is to align the spectrophotometer or spectrometer orthogonal to the tangent of the can diameter and perpendicular to the long axis of the can. If those angles are not maintained, then different reflectivities will produce slightly different measurements. The presently described embodiments incorporate several steps to solve these and other problems.

[0059] In the simplest form of the presently described embodiment, a spectrophotometer or spectrometer would be aimed at a fixed light spot that is somewhere within the field of view of one of the inspection system's digital cameras. The inspection system would have a priori knowledge of the exact aiming position of the spectrophotometer or spectrometer. The inspection system would also have knowledge and / or stored data about which areas of interest should have color measurements performed and would continuously monitor for instances where an area of ​​interest coincides with the position where the spectrophotometer or spectrometer had taken its most recent measurement. In other words, since the exact rotational position of the can is not necessarily known or could be random, the image captured by the camera would be used to determine the position of the decorated label. want Measure whether the light spot has In fact The measurement is taken from the most recent snapshot of the spectrophotometric data. A variety of techniques and resources can be used to achieve this, including the use of stored information, various coordinate information (e.g., XY or polar coordinates), and / or relationship information relative to the image. The logic and data used to track this can reside in the control of the machine vision system or in a separate or supervisory control. In either case, when the vision system determines that the spectrophotometer or spectrometer was truly aimed at the correct or desired area when it captured its inspection data, the vision system will instruct the system to accept the color measurement and will record the relevant data and information accordingly. If the correct light point was not measured due to the rotational orientation of the can, the data can be simply discarded or retained for use in other statistical processing.

[0060] For example, a vision system or control unit will have a database to store information, such as the information described above, including information about which positions on the beverage can will be measured and which colors are in the corresponding areas. The vision system or control unit will have knowledge of the absolute color specifications for each light point or area and / or store the absolute color specifications and will utilize the reported data of a spectrophotometer or spectrometer to monitor the light points or areas accordingly. In this simplest form, as previously described, the spectrophotometer or spectrometer will be aimed at a fixed point on the surface of the can. The rotational orientation relative to the long axis of the can is generally relatively random. Depending on the can handling system used and its position in the manufacturing line, there may be a slight tendency to favor a specific orientation; but in general, the orientation is reliably considered to be completely random. This means that any color measured at any given time using this configuration must fall within the field of view of the spectrophotometer or spectrometer by rotating the can only on its long axis. Thus, in our example, if all seven colors are brought into the field of view of the spectrophotometer or spectrometer by rotating the can only on its axis, then each of those colors will eventually appear and will be measured on a random basis in this inspection strip. If each successive can in a high-speed stream of container cans appears with equal probability in any given rotational orientation, then statistical randomness tells us that each of those colors will be measured with approximately the same frequency. If the colors that need to be measured on a can do not all appear at the same height, in other words, so that inspecting all consecutive cans with different rotations does not cause each of those cans to appear, then a more complex version of the presently described embodiment can be used.

[0061] It is entirely possible to implement the presently described embodiments so that the camera's image and the spectrophotometer's spot are acquired at different locations along the can's manufacturing travel path. As long as the can is not rotated between the image snapshot point and the spectrophotometer's data acquisition point, the same concepts taught herein apply. Of course, in this case, if the system does not provide illumination at the color reading spot, a spectrophotometer should be used instead of a spectrometer. Furthermore, if the material handling equipment (e.g., a vacuum wheel or star wheel) can ensure a known orientation relationship between those two key points, the system can be implemented accordingly to account for rotational position, if desired.

[0062] To incorporate more complex aspects of the presently described embodiment, a spectrophotometer or spectrometer can be mounted on a vertically adjustable mount that can be guided to different "Z" heights or vertical positions along the long vertical axis of the container. By doing so, color inspection of a completely new inspection strip at different "Z" heights around the perimeter of the decoration is facilitated. With the correct configuration, the vision system can then command the spectrophotometer or spectrometer to travel to different "Z" positions to begin random sampling within the inspection strip. Again, for the example of a seven-color can decoration, if the seven colors appear in an area of ​​sufficient size to completely fill the spectrophotometer's or spectrometer's field of view, then the spectrophotometer or spectrometer can inspect each of the seven colors after being sent to that can height strip position. If seven different "Z" strip positions are required in order to see all seven colors, then the spectrophotometer or spectrometer must be repositioned accordingly. But if, for example, those suitable color inspection areas only occur at three different belt levels, then the spectrophotometer or spectrometer would only need to be repositioned at those three levels so that it can inspect all seven colors at a location where it can view the appropriately sized area of ​​interest. The vision system or control system can then execute a program that repeatedly cycles the position of the spectrophotometer or spectrometer through a series of three positions. Depending on the relative importance of each of the corresponding colors, the dwell time at each belt level can be adjusted to ensure that a sufficient inspection frequency is achieved to correspond to the relative importance of each of the seven colors.

[0063] According to another aspect of the presently described embodiment, a spectrophotometer or spectrometer can be sent to the corresponding "Z" level inspection bands according to the user's inspection requirements and color specifications. If a particular color of interest occurs at five different "Z" band levels, the spectrophotometer or spectrometer can be sent to all of those levels or to a selected number of those levels on a sampling basis to ensure that those levels are monitored and measured according to the needs of the decorating process and the relative importance of the corresponding color zones.

[0064] For example, if a soft drink manufacturer uses three different colors to decorate a particular can, but one of those colors is a signature color, then one might want to focus most color inspection on the signature color. If the can's design causes the signature color to appear in all possible inspection band areas, then one might want to inspect for that signature color at all inspection band levels. However, this won't be possible if an area of ​​interest of sufficient size isn't available at each inspection band level to completely fill the spectrophotometer or spectrometer instrument's field of view. A repeating inspection band procedure can be used to, for example, achieve 14 of 18 feasible inspection bands. The priority and dwell time at each inspection band can be set to ensure that sampling is sufficient and statistically significant based on a percentage of sufficiently sized areas of interest around the band. There may be multiple possibilities around any given inspection band, and it may be feasible anywhere around the 360° perimeter of any given inspection band, or it may be feasible only at two locations around the 360° perimeter of the inspection band where there are sufficient areas of interest for color inspection.

[0065] The actual field of view of a spectrophotometer or spectrometer is entirely determined by the designer of the inspection system. There are many factors that will lead to the decision as to what the inspected area of ​​interest should be (including the relative complexity of the label design). Therefore, the expected size of the various areas of interest and the speed and focusing requirements of the selected spectrophotometer or spectrometer instrument design will all be factors in determining the field of view. Because the relative sensitivity of various types of spectrophotometers or spectrometers varies significantly, it may be desirable or necessary to have a larger or smaller field of view to allow for proper inspection and to collect enough light to work with the selected illumination system and technology. The use of a diffraction grating in a spectrophotometer or spectrometer inherently reduces the light available for integration at various wavelengths. Therefore, many spectrophotometers or spectrometers generally require more light than, for example, other sensors or even cameras.

[0066] In more complex forms of the mobile spectrophotometer or spectrometer version of the currently described embodiment, the vertical position and timing can be completely under programmable control. A closed-loop or open-loop servo system moves the spectrophotometer or spectrometer to the desired inspection belt height at the selected speed, according to the desired repeatability level, and according to the optimized pattern for inspecting all desired colors (according to their priority). However, the currently described embodiment can be practiced by utilizing any mechanical means that selectively moves the spectrophotometer or spectrometer to a given inspection belt "Z" position. Innovative implementers of the currently described embodiment can use any number of mechanical devices, including cams, linkages, solenoids, cylinders, or any other mechanical means that meets the cost and specification constraints of a particular embodiment. Vertical movement can be achieved manually using a simple hand crank or moving member, or simply slotted and manually adjustable to facilitate periodic adjustment to a single height.

[0067] An even more complex way to practice another aspect of the presently described embodiment is to utilize multiple spectrophotometers or spectrometers positioned at desired "Z" heights. For example, if two spectrophotometers or spectrometers are used, the number of vertical apertures or color inspection strips to which the spectrophotometers or spectrometers need to be sent can be reduced by approximately half. If there are twenty vertical inspection strips in a given application, and one spectrophotometer or spectrometer is located at the lowest or first strip and a second spectrophotometer or spectrometer is located at the tenth strip, only ten vertical movement positions will be required to obtain spectrophotometric color readings at all twenty inspection strips. To further this approach, a third, fourth, and up to N spectrophotometers or spectrometers with various vertical dimension offsets can be utilized accordingly. The system designer will have to determine the trade-offs for various system costs. The system designer will have to ask whether it is more expensive to have more spectrophotometers or spectrometers and shorter vertical motion requirements, or whether it is more valuable to have more spectrophotometers or spectrometers that can give more frequent statistical sampling of color at critical inspection bands. It should be readily understood that various trade-offs exist in system implementation to obtain the desired color inspection information at the desired statistical frequency.

[0068] In even more complex implementations of the currently described embodiment, a spectrophotometer or spectrometer can be deployed at each vertical inspection zone to eliminate the need to move the spectrophotometer or spectrometer group to any alternate vertical position. This, of course, saves the cost of moving mechanisms and associated controls, as well as the wear and tear on such equipment, and increases the frequency with which inspection samples can be obtained. As the cost and size of spectrophotometers or spectrometers continue to decrease, utilizing new, generally smaller technologies makes this implementation increasingly attractive and practical to completely eliminate moving parts. A single-chip spectrophotometer or spectrometer is small enough to be positioned in close proximity. It can even take the form of an array of spectrophotometers or spectrometers, spaced accordingly, and capable of simultaneously measuring the full color signature in a row of vertical measurement points. The same concept can be extended around the tank, allowing multiple points to be inspected simultaneously in any one zone. There can be one spectrophotometer or spectrometer for each camera, with multiple spectrophotometers or spectrometers in different zones within the camera's field of view.

[0069] Yet another aspect of the presently described embodiments having more than one spectrophotometer or spectrometer is that they provide some built-in redundancy. For example, it is possible to move two different spectrophotometers or spectrometers to the same inspection strip to verify the color results currently being obtained in that inspection location. This also means that a spectrophotometer or spectrometer sensor failure can have a smaller impact, as the other spectrophotometers or spectrometers can automatically be put into operation to perform the inspection strip color check for the failed sensor.

[0070] Another aspect of the presently described embodiments is verifying the color reproduction produced by the inspection system's color cameras. A spectrophotometer or spectrometer (a more definitive instrument) can actually be used to dynamically recalibrate the accuracy and absolute value of the information from the camera. The spectrophotometer or spectrometer can be used to electronically recalibrate the actual camera output on a periodic basis. Alternatively, the spectrophotometer or spectrometer can be used to modify the data in the vision system to achieve better accuracy and repeatability over time. Because the camera and vision system ostensibly see all pixels or areas of the can during each inspection, it has a significant advantage over a spectrophotometer or spectrometer, which only samples and provides usable data when oriented correctly. While any or all of the cameras can be used to determine the orientation of the decoration, and therefore, when spectrophotometer or spectrometer information is used correctly, the absolute color information can be used by one, some, or all of the cameras as a "standard" for calibration and recalibration. However, this can result in very robust data cross-utilization between a vision system connected to multiple cameras and one or more spectrophotometers or spectrometers.

[0071] For example, within a given inspection strip, a spectrophotometer or spectrometer connected to a control system or vision processor system can accumulate a statistical control chart for specific colors appearing at specific locations within the inspection strip. The statistical control chart can record L, a, and b color information and plot ΔE values ​​over time. The statistical control chart can also track the statistical mean and standard deviation for each measured zone. This can be done for each machine section (i.e., mandrel printing blanket, etc.) that contributes to the decoration of the can. The vision system can have all the statistical information about how the can looks, not only in the area of ​​interest of the spectrophotometer or spectrometer, but also in all areas of each inspected can. This data can then be compared and contrasted to determine a wealth of information about the robustness of the decoration process and various aspects of color printing.

[0072] Thus, in view of the above disclosure, the presently described embodiments, in at least one form, are implemented as a system and / or method (including a spectrophotometer or spectrometer and an image-based machine vision inspection system) for monitoring and controlling color decoration on cylindrical objects (e.g., rotating, non-oriented objects). The inspection system, in at least one form, includes at least one camera, an illumination system, a microprocessor or computer-based processor, and timing / triggering / control electronics.

[0073] The image-based monitoring inspection system is configured to capture an image of the perimeter of the color-printed cylindrical object in a field of view and feed the image data to a processor. The spectrophotometer or spectrometer is configured to obtain a light point color reading within the field of view where the camera of the image processing monitoring system is capturing an image and feed the light point color data to the processor.

[0074] Furthermore, the processor is configured to: utilize a priori knowledge of the coordinates of the locations where the light point color readings are located within the field of view of the camera to use at least one image logic algorithm to determine the exact position orientation of the color decoration on the cylindrical object imaged in the field of view, use logic to determine whether the light point color data is from the desired location within the image of the decorated label, use logic or rules to determine whether the light point color data can be used to compare with the actual orientation of the sampled captured image data, and output high precision color data based on the image data and the light point color data and the determination of at least one of monitoring the manufacturing process and correcting the manufacturing process.

[0075] The a priori knowledge of the coordinates of the light point color reading is determined by equipping the spectrophotometer or spectrometer lens arrangement with a laser that creates a recognizable mark on the camera's image. Furthermore, the processor has logic to locate the recognizable mark and record the corresponding coordinates for future verification when the laser is not turned on.

[0076] In these exemplary embodiments, the processor uses the spot color data from the spectrophotometer or spectrometer to perform at least one of verifying camera color information, recalibrating the camera, and augmenting the camera color output. In this regard, the processor uses additional logic algorithms to check calibration on all system cameras by statistically sampling and comparing recent color readings obtained by the spectrophotometer or spectrometer from non-simultaneous images formed by the camera, the non-simultaneous images including the spot region of interest of the spectrophotometer or spectrometer.

[0077] The processor also sends a signal to move the spectrophotometer or spectrometer position aiming point parallel to the central rotational axis of the cylindrical object, so that a different vertical band is monitored as the cylindrical object reaches the inspection position at various rotational positions. Alternatively, an array of more than one spectrophotometer or spectrometer is connected to the processor, with each spectrophotometer or spectrometer in the array focused on a different inspection band on the cylindrical object. In another alternative, an array of spectrophotometers or spectrometers is provided, with each spectrophotometer or spectrometer in the array focused on a different inspection band around the circumference of the cylindrical object, so that multiple bands of interest can be monitored without moving the spectrophotometer or spectrometer from a fixed position. In another alternative, the array of spectrophotometers or spectrometers can be moved as a unified unit to cover the desired area of ​​the color decoration while maintaining a fixed spatial relationship between the spectrophotometers or spectrometers. In another alternative, the spectrophotometer or spectrometer position can be one of: manually, semi-automatically, and automatically moved to a position so that its focused light spot is in the field of view of a different camera.

[0078] Regarding the output, the high-precision color data is displayed so that it can be used to manually correct the decoration manufacturing process. Alternatively, the high-precision color data is output to implement automatic correction of the decoration manufacturing process. In this regard, the output is connected to a control system that can directly drive servo adjustments to at least one of the digital inking head, additional print heads, print wheel adjustment, and individual spindle adjustment arrangements.

[0079] In addition, the high-precision color data based on the spectrophotometer or spectrometer is processed and organized so that the system can be initialized to output the data in the most useful form of the data, which can be one of (for example) CIE XYZ, CIELab, CMYK and CIE RGB. Still further, the high-precision color information is processed and output to determine when there is a statistically significant difference between the color quality standards that have been input into the system and the most recent actual production color readings.

[0080] In addition, in view of the above disclosure, the presently described embodiments in at least one form are implemented as a system and / or method for checking the accuracy of a camera implemented in an overall system having a machine vision system (having at least one camera, for example, multiple cameras, and in at least one example, four (4) cameras), a spectrophotometer or spectrometer, and a processor, or calibrating the camera. Such a system or method is capable of capturing an image or the color of an image of at least one section of the decorated periphery of a container, collecting accurate color information from a desired area of ​​interest of the decorated periphery where the camera has captured the image using the spectrophotometer or spectrometer, comparing the color authenticity of the image with the color information collected by the spectrophotometer or spectrometer, and using the comparison to check the accuracy of the camera's color if the desired section from which the spectrophotometer or spectrometer collected color information is located in a desired location on the decorated label. For example, the obtained information can then be used to correct the calibration of the camera, modify the color output of the camera, recalibrate the color settings of the camera, or recalibrate multiple cameras. Moreover, as described above, the obtained information can be used to monitor or correct the color decoration process.

[0081] To further illustrate the presently described embodiments, Figures 1 to 7B An exemplary way of reducing the embodiments of the present invention to be practiced is shown in FIG. , which is a key to the numbering used in the figures. It should be understood that the technical descriptions, features and / or functionalities described above in the description of the various currently described embodiments can be selectively incorporated into the combined embodiment alone or in various combinations as appropriate, as will be understood by those skilled in the art. Figures 1 to 7B In the example embodiments described.

[0082] like Figure 1 As shown in FIG, an architecture for monitoring and controlling, for example, in-line container decoration specifications within a can manufacturing system or facility is shown. In this regard, a stream (e.g., a continuous stream) 41 of cylindrical articles or objects (e.g., rotating, non-oriented cylindrical objects) (e.g., colored decorated containers, including, for example, beverage or food cans (e.g., which may be two-piece beverage or food cans)) is used in this arrangement, along with a conveyor or conveyor frame 30, a camera 10, and a spectrophotometer or spectrometer 20. Furthermore, processing and control functionality 100 provides suitable processing, storage, and control for the elements that facilitate implementation of the presently described embodiments. For example, a processor 110, a storage element 120, and a timing / triggering / control module 130 are implemented in one form. Furthermore, a lighting system or element 140 is representatively shown.

[0083] It will be appreciated that these elements (e.g., processor 110, storage element 120, timing / triggering / control element 130, and illuminator element 140) can be part of an overall manufacturing system or can be part of a high-speed machine vision system that is integrated into or functions in conjunction with a manufacturing system or facility. Furthermore, it will be appreciated that the various elements within the intended arrangement can communicate with other elements or be operationally or physically connected to achieve the desired control and / or feedback functionality. For example, controller 130 is representatively shown as being operatively connected to processor 110, illumination system 140, and portions of detection system 17.

[0084] It should also be understood that the presently described embodiments can be implemented using a variety of configurations, including a variety of suitable hardware configurations and / or software routines in conjunction with appropriate memory and / or storage devices, such as storage element 120 (e.g., read-only memory ROM, random access memory RAM, cache memory, any non-transitory computer-readable medium, etc.), as will be understood by those skilled in the art. For example, control and / or timing / triggering software routines can be stored on suitable memory or storage devices (e.g., non-transitory computer-readable media, etc.) and run on a processor or processor element (e.g., a hardware processor or processor element) within the system. These routines, in at least one form, will control suitable hardware devices, such as a camera, spectrophotometer, or spectrometer, etc., to implement the presently described embodiments.

[0085] More specific references Figure 1 and 2In an exemplary mode of operation, when a continuous beverage can stream 41 continues downward along the conveyor 30 in the can manufacturing facility, the can sits on the conveyor belt 31 and passes through the decoration inspection system installed on the system mounting structure 5. Although the individual beverage cans in the stream can be manually placed in the inspection position fixture as may be done in an inspection laboratory, this scenario details a high-production system in which the entire continuous stream is automatically passed through the inspection system. When beverage cans or cylindrical articles or objects 40 continue to pass through the inspection station on the conveyor belt 31 supported by the conveyor or conveyor frame 30, an encoder or resolver (not shown) is typically utilized to track the conveyor belt, which facilitates tracking the movement, speed, and position of the beverage cans. Now, when the beverage can 40 passes through a partial detection sensor transmitter-receiver pair (17A (e.g., transmitter) and 17B (e.g., receiver)), the system registers that the beverage can 40 is approaching the inspection point and tracks the beverage can by means of an encoder, which clearly indicates that the conveyor belt moves. In order to make the inspection system do its best, the beverage can 40 should be centered along the center line 32 of the conveyor and belt. Now, when the encoder or resolver tracks the position of the beverage can and the beverage can arrives so that its center (e.g., the center point of the long symmetry axis of the can from (e.g.) a top view) is located at 42, the image should be immediately and usually simultaneously taken by a plurality of cameras (e.g., four (4) cameras, as shown in this example) 10 surrounding the beverage can 40 and each mounted on a camera mounting structure 6 or 6A. Each camera has a field of view 12 that encompasses the field of view required for inspection, but typically the field of view is wider than the diameter of the can. If the system is designed and set up correctly, the horizontal center line 13 of the field of view should be perpendicular to the tangent of the outer surface of the beverage can as much as possible. Although the field of view 12 of the camera is typically wider than the diameter of the can, the center part of the image or the inspection periphery 45 of the can is typically used for further processing. The lens (e.g., camera imaging lens) 11 is selected to produce an appropriate field of view 12 and the lens should be positioned so that the center line 13 of its field of view is perpendicular to the side wall of the beverage can 40.

[0086] While each of the visual inspection cameras 10 will capture an image of its respective tank view, one of the images will be particularly important to the presently described embodiments. Camera 10A (in this example, a camera in whose field of view a spectrophotometer takes readings) has it associated with a spectrophotometer or spectrometer 20, both of which are mounted on a camera and spectrophotometer or spectrometer mounting structure 6A. The spectrophotometer or spectrometer 20 has a lens or optics system 21 that facilitates integration of information within a sensory circle, measurement area, spot, circular area, circular spot field of view, or data spot 25 of the spectrophotometer or spectrometer on the outer surface of the tank. The centerline 23 of the field of view arranged by the lens system 21 should define a narrow width field of view (e.g., the spectrophotometer or spectrophotometer field of view) 24 that will focus on the spot or circular area 25 (see also, for example, FIG. 2A ). Figure 4 、 5A and 5B) and should have a tangent in a plane perpendicular to the conveyor and perpendicular to a centerline 23 along the vertical along the outer surface of the can. Of course, it is quite possible to implement this with a spectrophotometer or spectrometer associated with each camera or movable to service each camera, but configuring the spectrophotometer or spectrometer in this way would be more involved and more expensive.

[0087] The color information imaged from the circular spot field of view 25 back to the spectrophotometer or spectrometer by the lens system 21 is typically incident on an integrating sphere 22 (e.g., within the spectrophotometer or spectrometer), which mixes or homogenizes the color information to achieve consistency. The spectrophotometer or spectrometer then transmits the color information homogenized by the integrating sphere 22 to a sensor array by means of a diffraction grating. The sensor can be a line scan imaging or area array imaging chip or a discrete sensor array, but the exact external dimensions of the spectrophotometer or spectrometer are beyond the scope of this invention. As mentioned, many different types of spectrophotometers or spectrometers and many different construction designs are commercially available. However, the spectrophotometer or spectrometer must be fast enough, have sufficient sensitivity, and have sufficient resolution to facilitate the application at hand by utilizing the same illumination provided for the visual inspection portion of the system.

[0088] Quality control guidelines from a brand owner customer or manufacturing plant will typically dictate which exact areas of a beverage can must be inspected for precise color. Figure 4 A beverage can 40 is shown having a color decoration around its periphery, and a light spot 25 to be inspected for precise color is shown. According to the presently described embodiment, the light spot 25 (and other locations) can be defined in a variety of ways, including using coordinates (such as XY or polar coordinates) or using relationship information relative to an image, or in other suitable ways. Figure 3A series of continuous bands 44 are shown around the periphery or girth of the can in the decorated area 16 (e.g., which may be within the field of view of a camera) that define an area or inspection region that may need to be inspected by the camera. The camera 10A associated with the spectrophotometer or spectrometer 20 has captured the image in the image. Figure 5A 4. The image represented by the area of ​​interest 12A in FIG. The area of ​​interest 12A is a portion of the entire image or pseudo-printing blanket 46 which may be composed of multiple images. It will be appreciated that the other areas or fields of view 12B, 12C and 12D originate from cameras 10B, 10C and 10D respectively. The sensory circle 25 is located within the area of ​​interest 12A where the spectrophotometer or spectrometer has collected its color data. The vision system will analyze the image of the area of ​​interest 12A and will determine where the light spot 25 happens to fall in the random orientation of the can that was just imaged. If the orientation of the beverage can happens to coincide with having the light spot 25 in the correct radial position on the band 44 on which it is focused (e.g. Figure 5A ), the system will be able to, for example, use the color information from the spot readings to determine whether the color is acceptable and, if necessary, use that information to obtain feedback to correct or monitor color during the decoration process and / or compare the color determined by the camera with the color determined by the spectrophotometer or spectrometer. It is possible to increase the frequency of obtaining useful spectrophotometer or spectrometer readings to define a tolerance band slightly larger than the precise inspection spot area (if this is permitted by the quality standards). A detailed comparison of the two areas can then be performed. This determination will heavily influence the direction of the spectrophotometer or spectrometer readings, which have a much higher resolution than the tristimulus camera sensor, as already described.

[0089] If the data spot 25 of the spectrophotometer or spectrometer happens to fall in an area that does not have a uniform color to be checked (e.g. Figure 5B If the color data is not shown in the figure, then the color data cannot be used to verify the isolated color. Figure 5A The location where the data "blip" falls Figure 5B The concept can be observed by looking at the difference between the positions in the image and the color data. Situations such as 5B do not render the data completely useless. Based on the image from camera 10A, the exact orientation of the decoration will determine exactly what color information will be available within the spectrophotometer or spectrometer data spot 25 for that given orientation. The implementer or user of the presently described embodiments or system will have to make a technical decision as to whether the color information is useful in various non-uniform (non-single color) locations. Some decorative patterns may be suitable for "mixed color" situations, where color data from such areas can be used to great advantage. An example would be an area composed of small mixed color dots, resulting in a rainbow-like area.

[0090] Return Reference Figure 3 , a spectrophotometer or spectrometer vertical positioning device 50 is also shown. The vertical positioning device 50 includes a base 51, a servo motor 52, a ball screw 53, and an encoder 54, but it will be appreciated that the vertical positioning device 50 can take a variety of forms. As shown, according to implementation of the presently described embodiment, the vertical positioning device 50 can be used to vertically move the spectrophotometer or spectrometer 20 to obtain readings from various belts 44 circumferentially disposed around the beverage can 40 along the can centerline 43 (e.g., the centerline of the can's long axis of symmetry).

[0091] refer to Figure 6 , showing another variation of the presently described embodiment. More specifically, instead of Figure 3 and 4 , a plurality of spectrophotometers or spectrometers 20 may be mounted vertically, each with its lens or optics 21 aimed to obtain a reading from a corresponding belt 44 of the cans being inspected on the conveyor belt 31. Furthermore, in at least one embodiment, Figure 6 An illumination field 14 with an illumination angle 15 of approximately 45° is illustrated.

[0092] Ultimately, the ultimate goal of monitoring the absolute color of various areas of a beverage can is to more precisely control the manufacturing process. Understanding periodic color drift relative to the correct specification, as well as all the various color and decoration variations, ultimately helps close the loop and dynamically or proactively correct problems. Indeed, sufficient data and tracking can be used to understand and ultimately correct many root-cause variations in the decoration process. Due to the historical difficulty in obtaining real-time information, there is currently a lack of available data on many of these root causes. The currently described embodiments allow for the collection of valuable real-time information, enabling manual corrections to be made in an intelligent and timely manner. It is expected that machine correction of decorators will be performed manually on older machines, then gradually become semi-automated when existing machines are retrofitted, and then fully automated as complex modifications are implemented and new machines are designed to incorporate the described technology. Servo control of many adjustments, automatic or digital control of the inking heads, and automated ink handling will allow the use of the collected color information to fully automate the closed-loop process. Obtaining precise data on the exact color applied during the printing or decoration process is extremely valuable for any correction process to be implemented.

[0093] In this regard, in Figure 7A and 7B, an example method 800 for checking the absolute color of printed objects and performing calibration according to the presently described embodiments is shown. It should be appreciated that, as mentioned above, this example method, as well as other methods according to the presently described embodiments, can be implemented in a variety of ways, including using a suitable combination of hardware, software, and memory / storage devices, as will be understood by those skilled in the art. As shown, a can is transported into an inspection area (at 802). A partial detection sensor verifies the position and initiates the appropriate color inspection sequence (at 804). Inspection is then initiated (at 806).

[0094] In this regard, a visual camera image is taken or captured at 808 and spectrophotometer or spectrometer information about a fixed light spot reading is collected at 810. Next, a vision algorithm is run to locate and orient the image at 812. The processor then determines whether the spectrophotometer or spectrometer has measured its light spot in the desired area of ​​interest based on the orientation of the can decoration at 814.

[0095] If not, the color information is discarded or saved at the "region" color database and a determination is made as to whether further analysis is required (at 816). If not, the data is discarded (at 818). If further analysis is required, the data is used to perform an auxiliary color mixing analysis (at 820). Next, a decision is made regarding the "color mixing" drift relative to the specification (at 822) and the system stands by for any future auxiliary or color mixing analysis (at 824).

[0096] Referring back to decision 814, if the processor determines that the light spot is measured in the desired area of ​​interest, the processor records and compiles the color light spot information along with all relevant light spot data, including location, machine part, time, and all other relevant data at 830. Next, the processor compiles a statistically significant sampling of the corrected location light spot data for the desired inspection area at 832. The color information is then used to close the loop with the decorator by sending color correction information at 834.

[0097] A determination is then made as to whether more spot readings are to be obtained at this spot (at 836). If not, the spectrophotometer or spectrometer is moved to the location of the new spot for inspection (at 838) and the process returns to 802 (at 840).

[0098] If more spot readings are to be obtained, the color information is used to compare the color information from the camera (if it were measured in the calibration position) at 850. The camera's color trends are correlated with the trend data from the spectrophotometer or spectrometer at 852. The camera color calibration update information is then saved to the camera or the computer to which the camera is connected at 854.

[0099] A determination is made at 856 whether the camera calibration settings are updated in a timely manner. If not, the information is stored for future use and the system is on standby for future data at 858. If the update is timely, the computer / processor updates the camera calibration settings and prepares the camera for further inspection using the updated settings at 860. The computer / processor also reviews a statistical sample of specific tank orientations and corresponding color readings in the area of ​​interest of the other cameras at 862. A determination is made at 864 whether the most recent camera color readings are sufficiently consistent with each other.

[0100] If there is a reading discrepancy between the cameras, a determination is made as to which camera most closely agrees with the spectrophotometer or spectrometer readings (at 866). The comparative readings and data are used to determine whether calibration changes or information adjustments should be made (at 868). If the data indicates that a change in lighting color has occurred (at 870), the system acts accordingly. Furthermore, the comparative readings from the cameras are used to determine whether the spectrophotometer or spectrometer is still functioning correctly or needs to be recalibrated (at 872).

[0101] Although it is beyond the scope of this invention to describe the hundreds of ways in which the various components of a decorator can be automatically adjusted, many people understand how to adjust a decorator but lack the real-time color and spatial information to implement automatic corrections and process control. This only occurs at a rate too high in most beverage can factories to be accomplished manually. The currently described embodiments should facilitate key step functions by enabling the availability of accurate, real-time information that the decorating process can utilize to operate in a closed-loop manner and significantly optimize compared to what was available historically. Those skilled in the art should be able to infer and use the information taught herein to apply the technical concepts to improve and close the loop on a wide range of decoration and printing applications.

[0102] The exemplary embodiments have been described with reference to preferred embodiments. Obviously, modifications and alterations will occur to others upon reading and understanding the foregoing detailed description. It is intended that the exemplary embodiments be construed to include all such modifications and alterations insofar as they come within the scope of the appended claims or their equivalents.

Claims

1. A system for monitoring and controlling color decoration on a rotating non-oriented cylindrical object, comprising: Spectrophotometer or spectrometer; and, an image-based machine vision inspection system comprising at least one camera, a lighting system, a microprocessor or computer-based processor, and timing / triggering / control electronics, said image-based monitoring inspection system being configured to capture an image of the perimeter of a color-printed cylindrical object in a field of view and feed image data to said processor; wherein the spectrophotometer or spectrometer is configured to obtain light point color readings within the field of view where the image is being captured by a camera of an image processing monitoring system and feed light point color data to the processor; and, wherein the processor is configured to: The a priori knowledge of the coordinates of the location where the light point color reading is located within the field of view of the camera is used to perform the following operations: using at least one image logic algorithm to determine the exact positional orientation of the color decoration on the cylindrical object imaged in the field of view, Using logic to determine whether the spot color data is from a desired location within the image, using logic or rules to determine whether the spot color data is useful for comparison to the actual orientation of the sampled captured image data, and High-precision color data is output based on the image data, the light point color data, and a determination of at least one of monitoring a manufacturing process and correcting the manufacturing process.

2. The system of claim 1, wherein the cylindrical object is a colored decorated container.

3. The system of claim 1 or claim 2, wherein the cylindrical objects are a stream of non-rotating oriented, color-decorated 2-piece beverage or food cans in the manufacturing process.

4. A system according to claim 1 or claim 2, wherein the a priori knowledge of the coordinates of the color readings of the light points of the spectrophotometer or spectrometer is determined by equipping the spectrophotometer or spectrometer lens arrangement with a laser, the laser making a recognizable mark on the image of the camera.

5. The system of claim 4, wherein the processor has logic to locate the identifiable mark and record corresponding coordinates for future verification when the laser is not turned on.

6. The system of claim 1 or claim 2, wherein the processor uses the light point color data from the spectrophotometer or spectrometer to perform at least one of the following operations: verify camera color information, recalibrate the camera, and amplify camera color output.

7. The system of claim 6 , wherein the processor uses additional logic algorithms to check calibration on all system cameras by statistically sampling and comparing recent color readings obtained by the spectrophotometer or spectrometer from non-simultaneous images formed by the cameras, the non-simultaneous images containing a spots-of-interest region of the spectrophotometer or spectrometer.

8. A system according to claim 1 or claim 2, wherein the processor sends a signal to move the spectrophotometer or spectrometer position aiming point parallel to the central rotational axis of the cylindrical object so that another vertical band is monitored when the cylindrical object reaches the inspection position at various rotational positions.

9. A system according to claim 1 or claim 2, wherein an array of more than one spectrophotometer or spectrometer is connected to the processor, each spectrophotometer or spectrometer in the array focusing on a different inspection zone on the cylindrical object.

10. A system according to claim 9, wherein an array of spectrophotometers or spectrometers is provided, each spectrophotometer or spectrometer in the array is focused at a different inspection band around the circumference of the cylindrical object, so that multiple bands of interest can be monitored without moving the spectrophotometer or spectrometer from a fixed position.

11. The system of claim 9, wherein the array of spectrophotometers or spectrometers can be moved as a unified unit so as to cover a desired area of ​​the color decoration but with the spatial relationship between the spectrophotometers or spectrometers kept fixed.

12. The system of claim 8, wherein the spectrophotometer or spectrometer position can be one of: manually, semi-automatically, and automatically moved to a position so that its measurement light spot is in a different part of the camera's field of view or in a different band of the cylindrical object.

13. A system according to claim 1 or claim 2, wherein the high precision colour data is displayed such that it can be used to manually correct a decoration manufacturing process.

14. The system of claim 1 or claim 2, wherein the high-precision color data is output to effectuate automatic correction of a decoration manufacturing process.

15. The system of claim 1 or claim 2, wherein the output is connected to a control system that can directly drive servo adjustments to at least one of a digital inking head, additional print heads, print wheel adjustment, and individual spindle adjustment arrangements.

16. A system according to claim 1 or claim 2, wherein the high precision color data is processed and organized so that the system can be initialized to output the data in the most useful form of the data, which can be one of CIE XYZ, CIE Lab, CMYK and CIERGB.

17. The system of claim 1, wherein the high precision color data is processed and output to determine when there is a statistically significant difference between color quality standards that have been input into the system and recent actual production color readings.

18. A method for monitoring and controlling color decoration on cylindrical objects in a system having a spectrophotometer or spectrometer and an image-based machine vision inspection system, the image-based machine vision inspection system comprising at least one camera, a lighting system, a microprocessor or computer-based processor, and timing / triggering / control electronics, the method comprising: capturing an image of a perimeter of a color-printed cylindrical object in a field of view; feeding image data to the processor; obtaining a point color spectrophotometer or spectrometer reading within the field of view where the image is being captured by a camera of an image processing monitoring system; feeding light point color data from the spectrophotometer or spectrometer to the processor; determining the exact positional orientation of the color decoration on the cylindrical object imaged in the field of view, determining whether the spot color data is from a desired location for a decorated label within the image, determining whether the light point color data is usable for comparison with the image data, and, High-precision color data is output based on the image data, the light point color data, and a determination of at least one of monitoring a manufacturing process and correcting a decoration process.

19. A system for color inspection of a generally cylindrical container having color decoration around its periphery, the system comprising: a machine vision inspection system having at least one imaging camera configured to capture an image of at least a section of the decorated periphery of the container; a spectrophotometer or spectrometer configured to collect precise color information from a desired area of ​​interest of the decorated periphery of which the camera has captured an image; at least one processor configured to compare the color authenticity of the image with the color information collected by the spectrophotometer or spectrometer, and If the desired section from which the spectrophotometer or spectrometer collects color information is located in the desired position on the decorated label, the comparison is used to check the accuracy of the camera's color.

20. A method of color inspecting a generally cylindrical container having color decoration around its periphery, the method comprising: capturing a color image of at least a section of the decorated periphery of the container; utilizing a spectrophotometer or spectrometer to collect precise color information from desired sections of the decorated periphery where the camera has captured an image; comparing the color authenticity of the image with the color information collected by the spectrophotometer or spectrometer; and If the desired segment from which the spectrophotometer or spectrometer collects color information is in the desired location, the comparison is used to check the accuracy of the camera's color.

21. The method of claim 20, further comprising at least one of the following: Using the color comparison information to perform one of correcting calibration of the camera and modifying camera color output; and Using high resolution spectrophotometer or spectrometer color information to at least one of monitor and correct the color decoration process; and wherein the generally cylindrical container is rotationally non-oriented.

22. The method of claim 20, wherein color settings of the camera are recalibrated as a result of the comparison between the color authenticity of the image and the color information collected by the spectrophotometer or spectrometer.

23. The method of claim 22, wherein other cameras in a multi-camera cylindrical container inspection system are also recalibrated, Wherein the computer / processor statistically reviews the camera color readings in the same trim location to determine possible recalibration settings.

Images