User:Spanishgeoff
Machine Vision Cameras
What is a Machine Vision Camera
Machine vision cameras refer to any camera that is suitable for use in industrial applications were the images of the objects being viewed are being processed immediately by an image processing programme which can be either be embedded in the camera or in a computer separate to the camera.
Different Types of Sensors used in Machine Vision Cameras
Machine vision cameras can contain either area scan or line scan sensors. The area scan sensors are progressive scan sensors and comprise of a 2D array of pixels. The number of pixels ranges from conventional CCIR sensor formats to many Megapixels. Generally the lower number of pixels the higher the frame rate. Cameras produced with this type of sensor capture each image as a whole frame therefore preventing image blur which is evident in the traditional interlaced camera. Linescan sensors are used in cameras which need to monitor very fast moving objects. (Example paper or textiles which may travel at many tens of meters a second) Such applications demand cameras that can read line data very quickly and the information can be processed on a line by line basis as the information is being transferred to the image processor. The camera also offers high resolution (several thousand pixels per line) as many of the objects being scanned can be very wide. Linescan cameras read data at many thousands of lines per second so can deal with defect detection in very fast moving objects. Machine vision cameras can use either colour or monochrome sensors. Colour sensors are monochrome sensors with a matrix colour filter across the front face. This will degrade the sensitivity but this can be compensated for by increasing the ambient light around the subject or lowering the iris setting to let in more light into the optic. Machine Vision cameras traditionally used CCD Sensors but in recent years cameras have become available using CMOS sensor technology. CCD sensors are more sensitive than CMOS sensors because the CCD sensor has 100% fill factor where the CMOS is only up to 70%. The fill factor relates to the active area of the sensor chip. The CCD sensor also has much lower noise characteristics than the CMOS sensor. However CMOS has the advantage of being more flexible than CCD in as much as the CMOS sensors can be windowed to read out a smaller amount of data at high frame rates, Also CMOS sensors have much lower power consumption.
Machine Vision camera output types
Machine Vision cameras are equipped with many different types of output offering high speed data transfer. For those cameras with an embedded image processor many pre-processing functions can be implemented in the camera before the data is transferred to an external computer for image processing. In recent years many different types of output have been used in the construction of Machine Vision cameras. Fire Wire IEEE1394 has been offered as a low cost “plug and play” so simple industrial processing. However, most serious machine vision applications now either employ the “Camera Link” the GigE Gigabit Ethernet output. The advantage of the later is that huge amounts of data can be transferred quickly and over long distances
Machine Vision Camera Optics
The optics for machine vision cameras needs special attention. Most optical systems in machine vision applications use fixed optics (i.e.: fixed for focus and iris). The application of machine vision has caused a new series of lenses to be produced specific for this application.
See also
External Links
Advice guide on how to select the right type of camera
United Kingdom Industrial Vision Association
British Machine Vision Association
European Machine Vision Association Spanishgeoff (talk) 10:57, 24 October 2008 (UTC)