Vision Systems

Adopting a vision system makes it possible to perform 100% inline inspection automatically for items that would otherwise need to be inspected manually. No additional costs are needed after the vision system is installed, and variation in judgment results between operators can be eliminated.

Improvements in vision system technology have enabled advanced, automated inline inspections for any level of user. It’s now possible to inspect 100% of products on high-speed manufacturing lines, so defective parts can be identified before additional processing or release. If similar inspections were done manually by operators, this process would be slower and less reliable.

Thanks to high-speed cameras and image processors, multiple images can be taken under different lighting conditions fast enough for 100% inline inspection. This makes it possible to complete multiple inspections at the same time, using variable lighting to pull out complex or low-contrast features and defects.

Using a vision system to act as the "eyes" of an industrial robot can significantly improve the accuracy and efficiency of advanced picking operations. The vision system detects a product’s position in the robot's coordinate system, and outputs these measurements directly to the robot controller for dynamic adjustment.

Conventional industrial robots require operators to specify coordinates manually using a teaching pendant. This manual operation can be time consuming, and accuracy may vary between operators. KEYENCE vision systems can communicate directly with many robot manufacturers, allowing for easy connection and efficient programming of the robot. This connection automates the cumbersome tasks of calibration and calculation, which stabilizes inspections and reduces the time required for integration.

In the automotive industry, even a small defect can result in a serious accident. To mitigate this potential risk, strict inspection requirements ensure rigorous quality and safety standards. Vision systems allow users to meet quality specifications, increase efficiency, reduce costs, improve accuracy, and ensure traceability of component inspections. Application examples include:

- Presence detection: moisture-proofing agents on ECU's, fuse assembly checks, and
- Appearance inspections: DPF, oil seal, engine valve, piston coating, and differential gears
- Measurement/alignment checks: battery positioning, bent connector terminals, spark plug dimensions, gather positional data for robotic pick and place applications

In the food and medical industries, inspection standards are regularly strengthened, and traceability is required to ensure safety. A vision system makes it possible to automate inspections and easily store image/result data. Application examples include:

- Presence detection: ensure proper case count, multiple components such as straws adhered to drink cartons
- Appearance inspections: character recognition (OCR) for lot codes, proper forming and no defects in food trays, shrink wrap, blister packs, and cans
- Measurement/alignment checks: label placement, seal width and position, needle dimensions, gather positional data for robotic pick and place applications

As electronics such as smartphones, gaming consoles, and PC's become smaller and thinner, their semiconductors and other electronic components need to be even more compact and accurate. KEYENCE vision systems improve inspection accuracy to meet the growing needs of these precise assemblies. Additionally, 3D vision systems bring ultimate stability to these inspections to identify height changes despite low contrast in materials. Application examples include:

- Presence detection: pins, connectors, solder
- Appearance inspections: crystal oscillators, IC molds, LEDs
- Measurement/alignment checks: Connector pin co-planarity, tray orientation, PCB warpage, terminal heights

Vision systems are used in the molding and resin industries to automate inspections and ensure defective products are not released. These inspections are necessary to detect any mishaps in processes such as injection molding, plastic forming, laminations, and product labeling. Application examples include:

- Presence detection: resin coatings, cap presence, proper labeling
- Appearance inspections: container inner surface inspection, product flash, resin scratching/chipping, foreign particles, pin holes
- Measurement/alignment checks: gasket center misalignment detection, label misalignment detection, bottle opening dimensions

A color camera is generally preferred when inspecting for changes or irregularities in color. Each pixel of a color image holds RGB information, which is three times the data of a monochrome pixel. Because there is more data per pixel, extraction and differentiation easier.
Monochrome cameras are preferred for measurement type inspections, where strong edge extraction is needed. Monochrome cameras are also commonly paired with colored lighting to aid in inspection. Examples include mitigating ambient light, causing a UV-dyed material to fluoresce, and emphasizing surface scratches.

Traditionally, colored lighting is paired with a monochrome camera, and the color (wavelength) of light will vary depending on the desired inspection. Colors which closely match a given target or are complementary to the target color can strategically manipulate and stabilize the image. To differentiate subtle differences in surface characteristics and further stabilize inspections, an intelligent lighting method called Multi-Spectrum uses eight different colors of light to gather more pixel data per pixel. Having data from multiple wavelengths creates more contrast between like-colors, and a true color image can be shown on the operation screen despite using a monochrome camera.

KEYENCE vision systems can perform appearance inspections on a variety of target surfaces. Sophisticated inspection algorithms, such as the defect tool, can detect localized changes in contrast to identify scratches and stains. Because the inspection looks for localized changes, external influences like ambient light do not disrupt the inspection. In addition to robust inspection tools, 24 image enhancement filters are available to reduce the effects of uneven lighting, rough surfaces, or variation among products. Two noteworthy filters are the scratch defect extraction, which enhances linear flaws on rough targets, and the shading correction, which emphasizes sharp changes in contrast while eliminating gradual ones.

Stable detection of burrs and chips relies on the accuracy of edge extraction. KEYENCE vision systems include profile defect tools, which precisely extract lines, circles, ovals, and free form curves to detect any sections which stray too far from the expected shape. The user can specify thresholds to control the size and severity of considered flaws.

All KEYENCE vision systems include a wide range of measurement and dimension tools, which make finding intersections, midpoints, and distances very simple. For example, simply click anywhere along an edge to auto-extract that line, or refer to line results from a tool which has already been configured. This intuitive interface allows users to easily create complex inspection settings while combining several measurements and dimensions.