In this article we look at two examples of how an emerging technology, intelligent lighting, is impacting applications along the lines highlighted by the VDMA focus group:
The concept of intelligent lighting has recently been realised through the adoption of industry standard networking and collaboration from leading manufacturers of LED lighting, LED control and imaging software. It’s an important development as lighting is a fundamental foundation for optimising inspection results and efficiency with machine vision - because even the best cameras can only capture what they can detect, and the best image processing software relies on these results from the cameras. The quality of illumination consistency, intensity and resolution will therefore always have fundamental consequences for users’ applications as explained below.
Intelligent lighting enables guaranteed consistent illumination levels over time by automatically adjusting lighting - based on image quality. This is useful in situations where the light degrades over time, or there are ambient light changes, or changes in the surface being inspected.
Currently, there is the option for adjusting camera gain, exposure and iris to compensate for inconsistent lighting within vision applications, but there are occasions where varying the intensity of the light has advantages over altering the camera or lens alone. If the gain is increased to cope with less light then the user will start to introduce noise into the image; increasing exposure can introduce blurring if the subject is moving; and opening the iris will reduce the depth of field, so if the part being inspected is not flat then parts of the image may no longer be in focus.
In these situations, it can be preferable to adjust the level of lighting to automatically compensate for any changes within the application’s illumination. With the Gardasoft Triniti intelligent lighting platform for example, the illumination intensity and quality for user applications can be guaranteed. This is because there is a seamless link between imaging software, cameras, lights and lighting controllers - to the point that the current illumination level at any time is known within the system, and the maximum illumination potential is also identified and levels can therefore be automatically adjusted according to the application’s requirements.
The diagram below illustrates this. The system has the following attributes:
In this example there is a GigE Vision camera and a Triniti-enabled light. In addition to its designated function, the camera also measures a defined area of the scene for illumination intensity - effectively a “test card” area of interest. This is continuously reported to the image processing application, where it is compared to the predetermined acceptable bandwidth of illumination for this particular inspection. If the bandwidth is contravened, then a signal is sent to the lighting controller to increase the lighting intensity; the controller will then increase the intensity until illumination levels are back within the acceptable application bandwidths.
Because of the Triniti system, the intensity of the light can be increased in a completely protected way because the controller is aware of the maximum intensity characteristic of the light. So the continuous cycle described for this application could be continued until the maximum light intensity had been reached. Note that the lighting controller would typically be configured to send a “preventive maintenance” warning signal to replace the light a while before any maximum illumination figure had ever been met (therefore increasing uptime of the manufacturing process and also eliminating unnecessary rejects due to poor inspection images).
Intelligent lighting also makes it easy to design and set up systems with what could be regarded as complex lighting configurations - for example, systems with multiple lights and cameras where many images are required of the same object. In these cases, benefits are realised because:
For example multiple images may be taken of a single product for different inspections. A ring light might be turned on full brightness for dimension measurements, then turned down lower to show surface defects, then the ring light turned off and a backlight turned on to check for other features such as holes being correctly drilled.
In the example below, the system is required to measure many different variables on individual products that are presented for inspection; it is also expected to handle this process for multiple product types. Multiple light settings are therefore required, and need to be applied each time a product type is identified at the infeed to the inspection process.
The lighting control function required here is called “multi-pulsing”, and this is available as an expert function called “S15” within the standard and Triniti versions of the Gardasoft LED controller family. The intelligent lighting platform makes this even easier to implement for ‘non-expert’ users because the lights are very simply identified, can be easily integrated with cameras, and can be commissioned in a very quick way via standard imaging software.
The controllers output a series of pulses of different intensities on different channels on each trigger. When the sequence finishes it restarts from the beginning and repeats. The length of the sequence, the different intensities and the pulse width on each channel is configurable. One input is used to trigger the light pulses and another can be used to reset the sequence to ensure that the system starts in a synchronised state. An Ethernet command can also be used to reset the sequence.
In the example, the lights are sequenced and controlled to take four images of the PCB. The lighting controller is triggered once for each image to be taken. On each trigger a different combination of lights is pulsed to take four images with different lighting schemes. When the next PCB arrives, the sequence repeats.
In conclusion, intelligent lighting is a good example of how vision technology is moving in response to expert user and application demand with enhanced control functions and seamless connectivity, but also offers an ease of use and intuitiveness which is encouraging new machine vision users.
Paul Downey, Marketing Manager, Gardasoft Vision Limited