- August 25, 2016
By Peter Bhagat, Gardasoft Vision
The ability of industrial vision systems to monitor, regulate, check, analyze, sort and classify with high precision, repeatability and safety makes vision an essential enabling technology in a host of automated manufacturing processes.
By Peter Bhagat, Co-Founder, Gardasoft Vision
The ability of industrial vision systems to monitor, regulate, check, analyze, sort and classify with high precision, repeatability and safety makes vision an essential enabling technology in a host of automated manufacturing processes. Vision systems are integrated into existing manufacturing lines, designed from the outset into new ones and frequently incorporated into OEM equipment such as packaging machines. For the automation industry, one of the key performance indicators is speed. Can we make product faster, better and at lower cost? Vision systems play an important role, and the primary focus from the machine vision angle lies with camera technology. We regularly hear about cameras with better resolution and faster frame rates capable of detecting smaller defects at faster speeds, but there are other considerations that are just as important in delivering these results. One of these hugely important factors is the lighting used for imaging, and every bit as important is the way in which that lighting controlled.
Why is lighting so important?
Machine vision does not examine the object itself - measurements are made on the image of the object on the sensor. Illumination is hugely important in the generation of an image with good enough quality for processing and measurement. The type, orientation and wavelength of the illumination can all play a crucial role in delivering good quality images, and LED illumination is the most commonly used in machine vision applications. However, whatever illumination geometry is required for a given application, the intensity of illumination is absolutely critical. If the resulting image has insufficient intensity, then it will appear dark, grainy (noisy) and not only might measurement accuracy be compromised, it might not be possible to make any measurements at all! Perhaps even worse, if there are fluctuations in the light intensity, variations between measurements lead to poor repeatability and that is simply unacceptable for an automated process. Applications such as high-speed imaging and multi-lighting schemes will require the illumination to be pulsed. Indeed the faster a production line runs, the camera needs more light for a shorter period of time so accurate control of pulsed lighting is of paramount importance.
Figure 1 Increased image intensity by pulsed overdriving of the LED compared to continuous illumination
Lighting control – getting the most from an LED
Although LED lights are specified by voltage, their light output is a direct function of current. In fact, a very small voltage change results in a large change in current, and consequently a large change in intensity. Thus all LED device manufacturers specify that current control be used to provide stable and highly repeatable light output. However there are many applications where running the light continuously, even at 100% output doesn’t deliver enough light for accurate measurements. One solution is to ‘overdrive’ the LED to boost the output (Figure 1). However the lights can only be overdriven in pulses to prevent damage. Control of the pulse width, frequency and intensity allows precise control of the overdriving and resulting light intensity (Figure 2). Overdriving LEDs makes it even more critical to use current control. Increasing the current by a given factor overdrives the LED by the same factor, but this level of accurate overdriving is not possible with a voltage-based driver. Converting a constant illumination system to pulse illumination is straightforward. The trigger for the camera is sent to a lighting controller that provides precise pulse width timing and brightness control for the lighting pulse. This ensures that the lighting is pulsed during the camera exposure time and the light energy is the same for every image. The basic principle of overdriving an LED to get more light out in short bursts is great, but care must be taken not to exceed the capabilities of the particular LED or there is a high risk that it could be damaged. In order to protect the light when overdriving, the lighting controller should be capable of regulating the duty cycle and pulse width according to the percentage overdriving required. In order to do this automatically, the controller needs to be supplied with the current rating for the particular LED. In addition, by continuously monitoring the current and voltage to the light it is possible to detect faults in the LED including integrity of connection, wiring faults LED failures etc.
Figure 2. Controlling pulse parameters to increase light intensity during cameras exposure time
High speed imaging
Another massive benefit of pulsing, or strobing the light source is to freeze the image of moving objects, particularly at high speeds, where overdriving may be essential to deliver sufficient light intensity for the short camera exposure times. This is perfect for automating vision inspection on busy production lines (Figures 3 and 4). A fine adjustment of the pulse timing in the lighting controller is often more flexible than adjusting the camera’s timing. The camera can be set for a longer exposure time and the light pulsed on for a short time to freeze the motion. The flexibility to choose the sequence for camera and lighting controller triggering means that a whole range of applications can be accommodated, with or without external triggers from product position sensors on the production line.
Figure 3. Pulsing illumination to ‘freeze’ movement on a fast moving bottle production line
Figure 4. Synchronizing illumination pulses to object movement
This versatile approach to camera and lighting controller triggering can further be extended to allow multi lighting schemes to be used. This may be required for the measurement of many different variables on individual products that can require the acquisition of a number of different images from the same part. These images could be acquired from a single camera or multiple cameras and are likely to require the triggering of a single or multiple lights at different intensities and for different durations in a defined sequence (Figure 5). Multi-lighting schemes can be used in a host of application types, including conveyor belt environments or in robot-guided inspections.
Figure 5. Example of multi-light, multi-camera inspection
So far we’ve considered the lighting controllers as separate (albeit essential) components of a vision system, but think how much more flexible things would be if the key elements of a vision system could be linked together. The intelligent lighting approach takes lighting control to a new level by the networking of LED lighting, camera and imaging software to provide an integrated application with a single graphical interface for set up and control. The intelligent lighting platform consists of three key elements:
- Light identification and operational data – special chips mounted into LED lights provide information on model data, electrical characteristics, optical characteristics and usage information for individual lights
- Integration of lights into software – these specially-enabled LED lights can be seamlessly integrated into machine vision networks via an SDK or through an API in major imaging processing packages. This provides users with a single graphical interface to set up the camera and lighting, visualize the timing and captured images, and save the settings to the camera and controller
- Expert light control- the systems incorporate all the functionality of LED light controller technology
This seamless link between imaging software, cameras, lights and lighting controllers means that the current illumination level within the system is known at any time. In addition, feedback from the lighting chip identifies the maximum permitted illumination allowing levels to be automatically adjusted according to the application’s requirements without damaging the light. In short, it provides non-expert users access to expert machine vision lighting techniques, enabling them to easily create, configure and commission vision systems with LED lighting.
Real time adjustment of light intensity
The use of intelligent lighting not only brings a holistic approach to the setting up of vision systems, but brings a number of other benefits as well. One of these is the ability to provide consistent light intensity over time. This is important in ensuring reliable, repeatable measurements, especially for applications where subtle variations in the image such as surface detail color variations or contrast levels are critical. There are a number of factors that can result in variations in illumination intensity. These include: age of the light, ambient light and other external factors such as dust, dirt and condensation, the temperature of the light, variations in lighting and camera exposure timing and variations in the drive to the light. Intelligent lighting provides a completely automated approach to constant lighting (Figure 6):
- The camera establishes the optimum illumination intensity for the application
- The image processing application constantly monitors acceptable operating illumination bandwidths
- The light controller automatically adjusts lighting intensity if necessary, in real-time
Figure 6. Closed loop lighting intensity adjustment based on monitoring intensity of a test image using the camera
The feedback provided by the special chip on the precise LED status means that the intensity can be increased in a completely protected way so this continuous cycle could be continued until the maximum light intensity had been reached. In addition, the lighting controller can be configured to send a “preventive maintenance” warning signal to replace the light at the appropriate time before any maximum illumination figure was reached (therefore increasing uptime of the manufacturing process and also eliminating unnecessary rejects due to poor inspection images).
Making it happen
Choosing the way that lighting is controlled in machine vision applications is just as important as choosing the best camera, lens, processing and measurement software and the lighting arrangement itself. The correct lighting control can make all the difference to a successful automation project using machine vision.
About the Author
Peter Bhagat is co-founder of Gardasoft Vision, with offices in Cambridge, UK and Weare, NH. Gardasoft produces high performance LED lighting controllers, including the Triniti® intelligent lighting system, and high intensity LED illuminators for the global Machine Vision and Intelligent Traffic (ANPR/ALPR)markets.Learn More
Did you enjoy this great article?
Check out our free e-newsletters to read more great articles..Subscribe