Technical Article

How is Industrial Machine Vision Transforming the Industry?

June 17, 2020 by Jeff Kerns

Liquid lens and multi-spectral lighting solutions are advancing the manufacturer’s production lines.

When people picture machine vision, they might think about Hollywood sci-fi movies with cyborgs or augmented reality in daily life. It might take time to see this vision become reality, but the building blocks of machine vision are being developed and are already changing the manufacturing industry. The following is a brief look at a few products that highlight how machine vision is accelerating production, increasing fidelity, and improving automation. 


machine vision

Figure 1. Machine vision is accelerating production, increasing fidelity, and improving auotmation. Image used courtesy of Cognex


Manufacturing is moving faster than ever with new challenges from short product life-cycles to mass customization. Technology is used to sense, document, analyze, and adjust production to keep up. This is all driven by data, which often starts with sensors. Perhaps the most data associated with a sensor is vision sensing. Each pixel of each frame is a piece of information and this can result in higher fidelity compared to other sensors.


Machine Vision - Companies and Product Types

According to a report that came out in December 2019, the main machine vision drivers are robotic vision, food packaging, and pharmaceuticals. The report also points out that machine vision, “... provides automated inspection and evaluation based on imaging for applications such as automatic inspection, process control, and robot guidance, typically in manufacturing.”

According to several sources, the key manufacturers in machine vision market include the following companies:

  • Cognex Corporation
  • Basler
  • Teledyne DALSA
  • Keyence
  • Datalogic
  • Edmund Optics
  • Allied Vision Technologies

Since there is a strong need for speed and quality, product trends in manufacturing are centered around getting the right image faster. 


Liquid Lens Technology

An increasing trend in manufacturing is flexibility, which means there might be multiple products or packages on a single production line. With multiple packages or tasks, a camera may need to refocus to scan labels, inspect products, or more.


Figure 2. Liquid Lens Technology. Gif used courtesy of Cognex 


To keep up with demand, cameras may need to process more than a thousand inspections or scans per minute. Fortunately, liquid lens technology is able to deliver a variable focal point and other benefits.

The benefits of liquid lenses include:

  • Less cameras used for applications requiring large depth-of-field
  • 2 to 3 milliseconds to refocus, ensuring consistently sharp images 
  • It’s rugged because it has no moving mechanical parts
  • It can operate in a variety of temperatures and requires little power

There are online tools to help determine which liquid lens might be right for different applications. There are many things to consider when selecting a camera, even one with a liquid lens. The focal range and how fast an image needs to be processed may determine the type of camera, lens, and lighting. 

For example, The following table shows the Field of View (FoV) widths of the 6.2 mm lens focused to 105 mm at various distances for the Cognex Dataman 260 (DM260) and for scale the last row is focused at 190mm.



Distance in mm


2D min. code


1D min. code


Area in mm






34 x 22






50 x 32


10 MIL




77 x 49


35 MIL


20 MIL


370 x 236

Table 1. Focused at 190 mm. Table from Cognex DM260 Reference Manual, courtesy of Cognex

Depending on the area and dimensions an application demands will determine the type of lens and operating distance of the camera. 


Machine Vision - Lighting

While a liquid lens can adjust to a range of distances in a few milliseconds, other considerations may affect the speed at which a device is able to process images. For example, using Power over Ethernet or a 5V USB power source for the Dataman’s illumination board yields a maximum duty cycle of 3% and a maximum exposure time of 0.5 milliseconds. However, powering the illumination board with a 24V external power source shows a maximum cycle time of 6% and a maximum exposure time of 1.0 millisecond.  

While liquid lens technology is keeping products in focus, lighting is still needed for the lens to ‘see.’ Changing colors, reflectivity, ambient interference, and more can make defect detection, alignment, or identification a challenge.

When it comes to machine vision, it can come down to wavelengths. For parts that have different or multiple colors to run on a single production or inspection line, having more than one type of wavelength is beneficial. For example, using a red light on a red part or label might make it difficult for a camera to detect or scan. Having multiple wavelengths in lighting solutions, such as Red, green, blue (RGB) lighting rings and bars, are becoming common in machine vision solutions. The right lighting solution might even save on the cost of a more expensive camera. 

For example, the Keyence Multi-Spectrum Vision System uses eight wavelengths for benefits such as the following:

  • Accurate sorting, even between slight color differences
  • Detect height changes while removing glare
  • Simple selection of optimal lighting

At an event, a Keyence representative demonstrated how a simple black and white camera, used with the Multi-Spectrum Vision System CV-X/XG-X Series, was able to produce detailed color images.


Figure 3. Keyence Multi-Spectrum Vision System Image Capture color analysis. Image used courtesy of Keyence

This type of cost savings and variability in products offering multiple focal points and lighting options are increasing manufacturing cycle times.

Liquid lens and multi-spectral lighting technology might help lead to more advanced machine vision seen in movies, but for now, it is meeting industry needs to keep supply chains and production lines moving faster without diminishing quality.   

*Article originally published on June 17, 2020, revisions have since been added