Contact image sensors realize efficient and economic on-line analysis The future of the broadloom inspection In the printing industry the demands regarding the product quality are constantly increasing. To ensure the desired quality in multi color printing, a reliable measurement and control system is required which verifies the web of the printed material for accuracy during the manufacturing process. Particular attention is necessary for color fidelity. A special challenge is the high speed with which the print product is transported over the web. In addition, the entire width of the print product must be mapped in equivalently high precision that will achieve high resolution and constant image sharpness without distortion. For broadloom inspection there are two analysis methods available: The use of line scan cameras or the installation of CIS contact image sensors. Following we will have a closer look at the operating mode, the pros and cons as well as the performance of both systems. The classic approach using area scan cameras and flash light sources has not been included due to restrictions such as distortion and problematic illumination when scanning broadloom goods. Broadloom inspection based on line scan cameras Currently, for the broadloom inspection of continuous print products most commonly a system consisting of line cameras in combination with a light source and lenses is used (Fig. 1, left). Fig. 1 left: Mechanical construction of a line scan camera, Fig.1 right: Mechanical construction of a contact image sensor
Like every other camera, the line scan camera captures images that are projected onto an internal sensor. From the sensor the information is passed to target systems such as display,, storage or analysis devices. In line scan cameras, the resolution specifies how many active pixels are located on the sensor altogether. This value with the reproduction scale of the lenses must be converted to an application-related measure. A pixel ( picture element ) is a two-dimensionally arranged and light-sensitive semiconductor element that serves for conversion of the incident photons the light - into electrons, buffer them and transport them forth the sensor surface. Most color line scan cameras produce color information by placing color filters above the pixels that lie on the sensor. Typically one red, green and blue sensor line is available on the chip (tri-linear camera). So called prism cameras, another type of line scan camera, produce color information by using a prism that splits white light into red, green and blue components for three different sensors (Fig. 2). As light source for line scan cameras, white LEDs are used. Fig. 2: Schematic representation of a prism camera For broad web inspection, several cameras must be placed side by side to cover the entire width. Undesirable spectral curve All types of line scan cameras have tremendous deficits in their spectral properties, caused by limitations, significant manufacturing variations of the filters and the deployed light source. The spectral sensitivity is the product of the light source and the sensor characteristics which means that deficits of the light source entail poor sensitivity. The spectrum of white LED s is composed of blue light from the LED itself and a yellow phosphor afterglow. Around 470 nm a spectral perception gap occurs, because at this point the blue
light wears off already before the effect of the phosphor begins (Fig. 3). Fig. 3: The spectrum of a white LED with a spectral gap at 470 nm These shortcomings can be mitigated by calculation with a color space conversion, but only at expense of the dynamics. The system behavior is a result of the light source spectrum combined with the spectral sensitivity of the sensor (Fig. 4). Fig 4: The range of on-chip filters Complex adjustment The installation of line scan cameras within the production line occurs in a certain distance from the web on a mechanical frame. When commissioning, a precise adjustment of the cameras is necessary to achieve correct results during the inspection process. Above all, the parallel connection of the cameras requires a considerable effort in the alignment and data processing so that a reasonably consistent image quality can be achieved over the entire area of analysis. If the production line changes and the print product is replaced a readjustment of all line scan cameras by specialists is required. The often long "levers" of the
image- and object size makes this solution sensitive to vibrations pixel sizes of a few µm cause significant effort in the assembly. Broadloom analysis based on CIS-contact image sensors The technology of contact image sensors has already been in a greatly simplified way established in fax machines and flatbed scanners. In industrial environments the Germanbased company Tichawa Vision GmbH, located in Friedberg near Augsburg, is the worldwide market and technology leader in the development and production of professional CIS contact image sensors. The Tichawa Vision CIS systems consist of a line detector with integrated lens and light source (Fig. 1, right). They work with a reproduction ratio of 1:1 therefore Tichawa Vision CIS systems are the same size as the target. In CIS sensors, the resolution is from the outset specified in dpi (dots per inch) or pixels per mm to suit the application. The Tichawa Vision CIS is installed with a working distance of 8 to 14 mm above the target object. The single lens is replaced by an array of staggered GRIN (gradient index) lenses (Fig. 5). Fig. 5: Cross-section of GRIN lenses Each lens captures a small area on the object and the controlled overlapping of the images produces a sharp image. The sensor types CCD or CMOS are used, installed offset or directly side by side.the essential application characteristics such as dark noise, peak response uniformity and dynamic range, the CIS meets the values of standard line scan cameras but without the usual compromises in spatial resolution and light output. Also, a single CIS can cover a web width of up to 4.1 meters (Fig. 6).
Fig.6: The boss delivers the goods in person! Mixed light source ensures optimum color fidelity The unsatisfactory spectral behavior of the line scan camera, mentioned above, does not apply to Tichawa s CIS systems. Contact image sensors are equipped with an optimized mix of different colored LEDs (Fig. 7). Fig. 7: The color spectrum of the CIS This prevents the emergence of a spectral gap and ensures an optimal spectral curve. The multiplex cycle with a single sensor guarantees a very precise geometric correlation of pixels of different colors (Fig. 8).
Fig. 8: The spectrum of an LED mixed light source Simple alignment and adjustment Contact image sensors are particularly suitable for use in applications with limited installation space. CIS solutions are placed at a distance of up to 14 mm from the web and are very compact in size. Contrary to the line scan camera, adjustment of the CIS is unnecessary. Optional ROIs (regions of interest) support the inspection of partial webs of varying width without readjustment. Constant luminous intensity and picture angle The CIS offers a complete solution including integrated illumination and lenses. Due to the 1:1 imaging, the mechanism is very simple. There are further advantages such as the constant light intensity,the constant focus and angle of view across the entire web width with a spatial resolution of up to 2,400 dpi corresponding to a pixel pitch of 10 µm. Color Fidelity An important factor for the quality of a print product is the color fidelity. Professionals often use the colors of the Munsell color chart as ideal values because the test with simple color charts or color coordinates in the color triangle is not sufficient for the high demands of the printing industry. Deviations from the Munsell reference values are undesirable. And of course, the inspection system that is used must approach the Munsell values as close as possible in its color reproduction. Munsell-Test: Line camera vs. CIS To determine which system provides better color accuracy, a Munsell test chart was captured by two different line cameras and a Tichawa Vision CIS. The obtained data were set in relation to the ideal values, as illustrated in the following graph (Fig. 9 and 10). The colored areas show the colors of the Munsell color spectrum. The white circle in the middle symbolizes the ideal results.
Fig. 9: Munsell color chart with shaded areas for the deviations of the line camera (left) Fig.10: Munsell color chart with shaded areas for the deviations of the CIS (right) The chart shows that over the entire color range, the CIS is much closer to the ideal values. With one-third of the area, the deviation of the CIS is considerably lower than the better of the two line scan solutions. Conclusion When comparing the two systems we must consider that the Tichawa Vision CIS provides a complete solution with integrated lens and light source, while with conventional line scan cameras, lenses, light source and fixturing must be purchased and mounted separately. Wiring is required for both systems. Regarding the investment costs for components and cables, the purchase of a Tichawa Vision CIS is recommended from a resolution of 2,000 pixels. If we also implicate the additional costs in terms of installation and adjustment, the Tichawa Vision CIS system results in a more economical solution at 1500 pixels web width. About Tichawa Vision: Since the founding in 1991 Tichawa Vision focuses on the development, manufacturing and distribution of camera technologies for industrial image processing for the purpose of surface inspection and product control. Tichawa is global market leader in the field of contact image sensors (CIS). Company Contact: Tichawa Vision GmbH, Burgwallstraße 14, 86316 Friedberg, Dr. Nikolaus Tichawa, Tel: +49 (0)821 455 553 0; Email: sales@tichawa.de Press Contact: epr-elsaesser public relations, Cornelie Elsässer, Tel: +49 - (0)821 4551 9178, ce@epr-online.de, Mona Moser, Tel: +49 (0)821 4551 9179, mm@epr-online.de, www.epr-online.de