In machine vision systems, acquiring images of moving targets is a challenge and consequently the best image requires three fundamentals to be well defined:

  • an excellent camera;
  • an appropriate lens;
  • an appropriate illumination.

All three items are the key to enable subsequent successful image analysis, while poor image quality will task any machine vision application.

Figure 1. Signal to noise ratio (SNR) related to illuminance.

The illumination choice is the first step that directly affects the quality of the images. There are many difficulties that can result from the wrong illumination selection. In general what is not illuminated correctly cannot be evaluated by software, or even by humans. A camera can also not be compared to the human eye that adapts automatically and is very flexible to difficult tasks. A human would even change the distance or the angle of view to discover details. For a machine vision camera it is not possible to detect edges if the object is not illuminated correctly.

In today’s vision applications, different tasks have to be solved. Metallic shiny, matt dark, or even transparent surfaces with different features require different types of illumination which led to the development of different illumination technologies. Many aspects affect the choice of correct illumination and have to be considered:

  • area to be illuminated;
  • camera in use;
  • speed of the application and the camera itself;
  • color of the illuminated objects;
  • environment;
  • behavior and characteristics of the object (glossy, diffuse, height variations,…);
  • expected / required lifetime of the application.

If all of these parameters are understood, the application can be simulated and it can be determined which illumination fits best.


Figure 2. Example of SNR resulting from shot noise // 20 μs integration time, 200 DPI resolution, diffuse reflection with 80%.

Many people are not aware of what the ingredients are for a good machine vision application. In line-scan applications the light should be where the sensor array(s) of the camera are focused. Light outside this area is wasted and results in extra costs and heat.

Figure 3. Spectral changes of an LED over temperature, 55°C as reference.

In all imaging technologies, one important quality criteria is noise. There are several sources of noise in an imaging system but normally, the shot noise dominates. Shot noise is caused by a physical effect and has nothing to do with camera quality. The reason for shot noise is found in the discrete nature of the light (photons) and the resulting discrete generation of electrons in a sensor pixel.

Shot noise has a Poisson distribution and therefore, the signal to noise ratio can be described as: SNR = √Ne

The number of electrons is directly proportional to the number of photons. The number of photons is directly proportional to the product of sensor illuminance and exposure time. In a given imaging setup with a defined optical transformation there are three parameters that influence the shot noise in an image:

  • integration time (scanning speed);
  • f-stop (depth of focus and maximum sharpness);
  • illuminance on the scanned object.

The f-stop of a lens has a significant impact on the requirements for light. For example, changing the f-stop from 4 to 5.6 increases the light requirement by a factor of two if trying to keep the same signal-to-noise ratio. At the same time it increases the depth of focus and improves the optical quality with most lenses. So the depth of focus and sharpness increase while vignetting effects are reduced. What machine vision application wouldn’t benefit from having a sharper image and an increased depth of field?


Object resolution: 300 DPI

Sensor pixel size: 10 μm

Focal length: 50 mm

  • Depth of focus: 8 mm // f-stop 4.0
  • Depth of focus: 12 mm // f-stop 5.6
  • Depth of focus: 18 mm // f-stop 8.0

General hints:

  • Increase the f-stop number and increase the illumination so that images are sharper while not reducing the signal to-noise ratio.
  • Increase the brightness to a range where the camera reaches 80% (or more) of the sensor saturation when scanning the brightest area of an object.
  • Shot noise is one of the most important parameters for image quality. Collecting as many photons as possible within the defined integration time will increase image quality.
Figure 4. Reflector focus of 190mm (left). Reflector focus of 95mm (right)

These conditions in image acquisition reduce risks, increase quality and reduce the processing power in PCs that are required to overcome problems during the image processing. The speed of line-scan cameras has increased significantly during recent years and it is becoming more and more of a challenge providing enough light for these fast systems. Anyone with a digital camera already realizes that poor light conditions lead to poor (noisy) images. The same happens in machine vision applications, so the challenge is to produce enough light for these difficult applications.

Figure 5. cos4 (Phi) related brightness decrease.

The importance of light for camera applications is shown by the effects of shot noise. Image quality is strongly correlated with the number of photons on the object and, ultimately, the sensor used for acquisition. Modern line scan cameras are able to deal with integration times down to e.g. 15 μs; a luminance of more than 1 Million Lux can be required to produce optimum image quality. The two images in Figure 1 show signal to noise problems that are caused by lack of light.

The chart in Figure 2 shows a typical setup with SNR related to illuminance.

Some lighting systems on the market provide focused light solutions to increase intensity at the point of imaging. When it is not a backlight or bright field illumination, it is a challenge generating enough light on the object so that line-scan cameras can provide perfect images. This is why rod lenses are commonly used to focus light. While rod lenses cause color aberrations resulting from diffraction effects, a mirror (reflector) principle is free of such issues. With reflector technology it is possible to collect more light (wider angle) from an LED, thereby increasing efficiency.

Lifetime and Degradation

Illumination technology today is constantly changing from classic lights like halogen or fluorescent bulb lamps to LED-based lights. The success of the LED is driven by a lot of advantages and industrial improvements. Classical light sources (e.g. halogen bulbs) change spectrally, take a long time to reach a stable state, and have a very limited lifetime. The spectral behavior of LEDs is stable when the temperature and current are held constant and the LED light sources are ready for operation almost immediately.

The short time for warm up is because of the small size. This is one reason why good thermal management is needed to keep the LEDs at a reasonable working temperature. The recent improvement in the stability of LEDs enables a long lifetime and a constant quality of spectral behavior. There are different options to increase the lifetime by adequate thermal management and also by controlling the operation mode of LED illumination. This means, for example, if LED illumination is strobed, which is technically easy to implement, the lifetime will increase. With strobing technology it is partially possible to use the LEDs over the maximum current for tasks where intense light is required. But be aware that the latest high-power LEDs do not support excessive currents as before.

The natural behavior of an LED over its lifetime is that the intensity will decrease. Measurements showed that in 50,000h the intensity decreases from 100% to 70%.


Is LED light really a cool light source? If LEDs are driven hard without cooling, they burn out and die within seconds. Cooling is an issue and the better the cooling, the longer the lifetime of the LED. This is now common knowledge, but there are other adverse effects to consider.

LED temperature influences:

  • Lifetime Spectral behavior / color shift (see Figure 3);
  • Efficiency / brightness.
Figure 6. Illumination length vs. distance to the object.

Figure 3 illustrates the color shift caused by a change in LED temperature. The listed values are the mean and maximum color distances in lab space, measured on a Color Checker test chart and referenced on the 55 °C-Values.

Figure 7. Spectral emission of a typical white LED. The dotted line shows the spectral sensitivity of a human eye.

As can be seen, spectrum changes due to temperature variation. 30°C difference can have an influence on the spectral behavior from 2. 2 to 12.4 ΔE (delta E).

(Remark: E>1 will result already in visible color changes.) In applications where precise color reproduction is essential, it is strongly recommended to guarantee the thermal management of the illumination. Active thermal control systems can control the LED’s temperature by intelligent cooling in a narrow range of less than 2 degrees. In general the following cooling options are available:

  • passive cooling by heat sink;
  • compressed air cooling;
  • liquid cooling;
  • fan cooling;
  • temperature controlled fan cooling.

Active air flow, compressed air, and water cooling are best for measurement applications in high temperature environments. By monitoring the temperature of the LEDs and controlling the cooling (e.g. fan-based) ΔE issues can be avoided. Cooling helps to minimize the color shift of the LEDs and leads to more accurate measurements. It also helps to increase the lifetime of the LED.

The Challenge of Different Working Distances

If an engineer has an application where the object is flat and the working distance is known, then there are fewer issues to solve. The selection of the focal length can be relatively easy in this case. But what to do with varying distances because the object does not have a well-defined distance to the light? Reflector technology enables a greater usable depth with an illumination source. The images in Figure 4 show the advantage of the reflector technology in regard to the distance. Homogeneity is needed over the length, width and depth.

With white LEDs spectral issues are commonly caused by chromatic lens errors. The best illumination solution focuses the emitted LED-light via the newest reflector technology. Focusing by reflector technology does not lead to chromatic aberrations that can be a challenging issue in applications with white light and varying distances. A focused light with no aberrations, even with different angles and distances, increases the efficiency and stability in applications.

If an object has varying distances to the camera/light, varying brightness is an additional challenge for a subsequent image processing task. So a careful selection of a light source that has minimal changes in brightness – even with varying distances – is highly recommended.

Choosing the Right Illumination Length

Figure 8. Images with and without polarizer.

Every lens causes a decrease in brightness, resulting in raw images when viewing angles are further from the center towards the edges. The decrease of brightness is proportional with cos4 (phi) where phi is the angle measured from the optical axis. As can be seen in Figure 5, there is a significant brightness decrease from the center to the outer regions of an image. So there are good reasons not to have wide viewing angles as the brightness and accordingly the SNR will decrease from the center to the edges.

Additionally the right length of a light is important for any application to avoid additional problems at the outer regions of images. While the center of an object receives energy from both sides, the outer edges of an object will suffer a lack of light.

It is recommended to have illumination modules that are longer than the object itself (Figure 6). The larger the distance between the illumination and object, the longer the unit should be. To have ample light conditions the following rule of thumb can be used: IL = FOV + 2*D

White LEDs

LEDs are available in many different colors like red, green, blue, yellow, amber, white, IR. UV LEDs are also available, but with wavelengths smaller than 365nm the lifetime is very short and emissions are weak. On the other hand IR- LEDs with wavelengths greater than 950nm have a very limited output. Nevertheless different colors and wavelengths help to make things visible on surfaces with different spectral behaviors. In the past red illumination was often used where high intensities were required. The latest performance boosts in LED technologies are mainly in white LEDs. We see these high-power LEDs in headlights of cars or even street lamps. The core of a white LED is in fact a blue LED. It is the phosphors that convert part of the blue to the remaining visible spectrum.

Figure 7 shows the blue peak near 440 nm of the blue LED inside a white LED. The remaining part of the spectrum results from the conversion material (phosphors). Technology-wise it is a challenge for LED manufacturers to keep the color of white LEDs stable. There are tolerances in the blue chip and additional tolerances in the conversion material. This all leads to unwanted variations in white colors produced within the same production line. Consequently, LED manufacturers classify LEDs into different groups. Each group (binning) has a certain tolerance range with respect to efficiency and color. This should be considered when selecting an illumination source. If the illumination modules differ in color from piece to piece, or even inside the same illumination unit, it will complicate the image analysis.

Non-Visible Spectra

UV LEDs are used to make fluorescence marks visible. In many cases 405nm light sources are sufficient to excite fluorescent substances. On the other hand, for curing processes of glue, varnish or resin, UV LEDs may be the better choice. UV LEDs are not as powerful – compared to blue/white LEDs – but the intensity can be increased by focusing the UV beams via reflector technology. IR lights are used in food inspection applications to detect organic materials. Wavelengths of 1200-1700nm can help to distinguish between different materials. Unfortunately today’s IR LEDs in these wavelength ranges are not yet powerful enough, so conventional halogen bulbs with filters are used.


Reflecting materials or surfaces are always a challenge with respect to illumination. In combination with a 90 degree rotated polarizer in front of the camera, unwanted light reflections can be eliminated (Figure 8).

Polarization in technical applications is a challenge. On one hand, it is key that the illumination module will stay within certain temperature ranges. On the other hand, a crossed polarizer arrangement cuts down the efficiency to approximately 18% of the initial emission. So with respect to a good SNR (signal-to-noise ratio), the primary emission needs to be very high, enabling acquisition of good images even with the crossed polarizer arrangement.

LED Controllers – What Are the Key Parameters?

There are different concepts on the market. Some controllers are integrated, while others are external, and sometimes not even used. However, an LED controller is not just a different kind of power supply; it can be the key to success.

A stable inspection unit requires a stable environment. If the LED controller is not stable in terms of temperature variations or supply current to the LEDs, brightness changes can make material inspection impossible. Especially with the very high frequencies of line-scan cameras, special care is required with the selection of the LED controller. Short-term fluctuations in brightness will immediately be visible.

In industrial environments, a robust design and the right choice of interfaces must be available. Control interfaces allow adjusting the light output remotely e.g. different materials need to be inspected on the same production line but require different light levels. Available interfaces are:

  • Ethernet;
  • RS485;
  • USB;
  • RS232;
  • PWM;
  • Analog.

If controllers are designed for special light systems, a monitoring /surveillance function can also be integrated. An inspection system can read the LED and controller temperature as well as other useful information. This enables stability analysis and control in a machine vision application.

For more information, Click Here