Machine Vision Applications Courtesey of Stemmer Imaging
The following applications indicate the diverse nature to which machine vision technology can be put.
- On-board And Trackside Railway Inspection
- Synergy between Solar cell and Machine vision Technologies
On-board And Trackside Railway Inspection
Cameras and image processing techniques originally developed for machine vision applications have been used for a host of railway inspection applications. Cameras of various types can be located on the train itself or by the side of the track, depending on the particular application. Railway inspection techniques present a significant challenge to the vision industry. Not only is a wide range of imaging devices, including area scan and high definition cameras, linescan cameras and thermal imaging cameras, required, the operational conditions are demanding. Cameras and imaging systems can be exposed to extremes of weather, vibration and physical wear. STEMMER IMAGING is well placed to meet these demands, with a comprehensive product range that covers the entire range of components and services that required to produce an imaging solution, combined with the experience of working on a number of railway projects throughout Europe.
Trackside applications
Trackside mounted cameras can carry out a variety of tasks. Cameras mounted on posts can be used, for example, for vehicle identification and to inspect passing trains for graffiti. It is important to identify whether the graffiti was applied while the train was in a depot, or while it was stopped elsewhere, since responsibility for the train lies with different organisations depending whether it is out on the track or in the depot. The origin of the graffiti, therefore, has implications as to who is responsible for cleaning. Other trackside applications include cameras mounted in troughs on the track pointing at the brake shoes which can be used to evaluate brake shoe wear or to examine the wheel profile for damage and wear which can be useful for predictive maintenance, since the wheels can damage the rails and the rails can damage the wheels!
Train mounted applications
Cameras can be mounted in a wide variety of locations on the train itself. Monochrome, colour and thermal cameras can all be mounted in a forward facing position (to give a driver’s eye view). General monochrome cameras give a good understanding of the items that are beside the line such as trackside cabinets and huts etc and so can be used for asset recording and mapping with accuracy to within a couple of metres. Using high quality 3CCD colour cameras in this position allows the visibility and quality of railway signals to be assessed. Thermal cameras can be used to check that heaters are working in trackside cabinets (which can be important for operation of the signals); that points heaters are working or to locate hotspots generated by the breakdown of insulators on the third rail. Forward facing cameras can also be mounted near the top of the train pointing downwards to give the ‘4ft view’. This replicates the view a track inspector would have and is used for checking ballast, missing clips, tie bars, inspecting points etc. Although the assessments are still made by a human inspector reviewing the video recorded by the cameras, the use of cameras avoids the need to close the track for the inspection, saving both time and money.
Roof-mounted monochrome cameras provide information on pantograph and catenary interactions, especially with regard to wear and contribute towards design improvements and preventative maintenance. Cameras mounted on the bottom of the train provide information on wheel-rail interfacing, contact point friction and the dynamics of conductor shoe movement on conductor rails. Laser triangulation techniques allow 3D measurement of rail gauge, including structured gauging and looking at the profile of rail heads at high speeds. Linescan cameras mounted in this position provide very high resolution track views and provide an alternative to the ‘4ft view’ for checking the presence of clips etc. This high quality sub millimetre scale imaging again removes the need to send a man out and close the line for inspection.
STEMMER IMAGING has been involved in a number of railway inspection projects, and while fully automated inspection is not yet an accepted methodology, there are some areas, such as track inspection, where the high degree of repeatability and quality of the images produced by automated imaging techniques are showing excellent results and are being used in conjunction with human inspectors. Although most of the inspection methods discussed still require human interpretation of the images produced, there is plenty of evidence that vision techniques can offer significant cost savings.
Synergy Between Solar Cell And Machine Vision Technologies
It is interesting how two technologies can be mutually beneficial. Machine vision inspection has been used to provide real-time process feedback in the manufacture of solar cells, while solar cell technology has been used to develop a machine vision sensor which offers exceptional dynamic range for demanding vision applications.
Improving solar cell manufacture
ECKELMANN AG, located in Wiesbaden, Germany has developed a vision system based on line scan cameras for laser edge detection as part of the edge isolation process in solar cell manufacture. It was designed for the ASYS Group (Dornstadt, Germany), a leading manufacturer of handling systems, process machines and special machines for the electronic and solar industries, and is fully integrated to provide feedback control to the production process. Edge isolation provides electrical separation between the active front side of a solar cell and the rear side. A laser cuts a small groove along the cell edges, the depth of the groove depending on the cell doping. The difficulty lies in positioning the groove as close as possible to the outer contour of the cell in order to maximize the active surface and thus the efficiency. The edge isolation control system features a line scan camera with 4096 pixels, optics and customized LED illumination supplied by STEMMER IMAGING. The image processing system measures the outer contours of the cell and feeds them back to the control system of the laser equipped with a deflection mirror to provide an active feedback system. If the edge damage is within tolerance levels the laser will ignore it and proceed with the cutting process. Image acquisition and analysis take place in just 800 ms and the resolution of the system makes it possible to ensure that the distance to the edge during laser cutting is below 100 µm. The calibration and qualification of the laser and camera have been automated so the system can easily be commissioned or recalibrated after maintenance work.
High Dynamic Range cameras based on solar cell technology
The new patented FX4 HDR sensor from German camera supplier IDS features miniaturized solar cells rather than the photodiodes used in conventional CCDs to produce extended dynamic range. High dynamic range (HDR) machine vision sensors aim to mimic the capability of the human eye to image details in scenes which contain both very bright and very dark areas. While the eye can perceive all brightness levels, conventional CCD sensors suffer from overexposure and therefore lose image data. HDR technology, on the other hand, enables fine differences in brightness to be imaged even in very bright scenes. A conventional image sensor with a dynamic range of 60 dB could image a scene dynamic of 1,000:1, i.e. the highest brightness value is 1,000 times brighter than the lowest brightness value. The human eye can perceive a dynamic range of up to 100 dB within a scene, which corresponds to a brightness ratio of 100,000:1. The new HDR sensor has a dynamic range of 120 dB, 1000 times greater brightness ratio than conventional CCD sensors. Traditional photodiode sensors generate a linear current proportional to the amount of light, while solar cells output a logarithmic voltage based on the amount of light falling on them. The logarithmic response of this sensor not only means that large changes in brightness in light areas of a scene cause only small changes in image brightness, but also prevents saturation in the image and ‘blooming’ (where charge from overexposed pixels ‘overflows’ into neighbouring pixels, causing whole areas of an image to appear white with the loss of image data). The new sensor does not use integration methods and so operates completely in real time. High dynamic range applications include automotive/traffic, welding, paints/glossy finishes and payment kiosks. For example, in traffic applications the sensor may need to identify detail in the dark interior of a vehicle while the headlights are on.
