This article describes the benefits of line scan cameras, including the ability to image large objects and the perfect, high resolution images.
A single line of pixels in a line scan camera images across an object. A two-dimensional image of the object is built up by moving either the object or camera perpendicular to the line of pixels. This may look like a complex way to image an object when compared to “frame cameras” that take two-dimensional images (a “frame”). However, when the task needs perfect or high resolution imaging or the object is large and continuously moving, then a line scan camera is usually a better choice compared to a frame camera.
If a machine vision system needs to inspect long rolls of material, such as plastic or paper, for any sign of defects, a resolution of 16,000 pixels across the material is required to locate these defects. The material is unrolled under the camera and using a rotary encoder attached to the unrolling mechanism, the speed of the material is measured.
While users can use four large-format frame cameras mounted side-by-side across the web of material, this has several drawbacks. First, large-format frame cameras often have defective pixels that are concealed by filling in a pixel value from adjacent pixels and this can conceal the defects users are attempting to detect. Second, the exposure time of the frame camera must be short to avoid blurring caused by the movement of the material. A short exposure needs high intensity lighting over the whole imaged area to achieve a satisfactory image. It is the rotary encoder output that triggers synchronous image acquisition for all cameras. Third, removing gain and offset variations in pixel responses – needed to detect low contrast defects – takes a lot of time with frame cameras.
A much better solution is a single, 16,000 pixel line scan camera. There is one camera that now spans across the entire material and there are also no defective pixels. The acquisition of each line of the image is triggered by the rotary encoder output. Motion blur is prevented with short line exposure time. However, the illumination can now be focused into a line, instead of being spread out over a frame, to obtain an acceptable image. A Time Delay and Integration (TDI) line scan camera features rows of pixels that integrate exposures in synchrony with the material’s motion, thereby multiplying the effective exposure time by the number of rows of pixels. Last, it is possible to adjust the offset and gain of each pixel to give an even response over the line of pixels, and these adjustments are made in the hardware of the line scan camera.
When to Use Line Scan Technology
Line scan technology is suitable for those applications where there is a need for large, high-resolution or high-speed image capture, such as in a continuous web application—textiles, paper, glass tape, or metal, for instance—or for road and railway inspection, or for surface topography scanning or satellite imaging. Line scan technology works suitably with free-falling products, including steel and molten glass or postal sorting. Line scan is also suited for various tasks such as inspection of flat panel displays, silicon wafers, printed circuit boards, solar cells, medical diagnostics as well as other tasks where large, defect-free images are required.
When the product speed is high, line scan cameras provide a better option than frame cameras. This is because short exposure times can be compensated by using Time Delay and Integration (TDI) and concentrated illumination to increase the photons “harvested” by the camera. While a line scan camera performs gain and offset correction in the camera, an area scan camera uses the CPU of the vision system to calculate this correction and therefore is much slower.
How Line Scan Images are Formed
One or more lines of pixel sensors are present in a line scan sensor. During the camera’s exposure time, each pixel builds up photoelectric charges that are proportional to the light from the object imaged onto that pixel. Towards the end of the exposure time, the charges in an entire row of pixels are transferred into a readout register. These pixel charges are shifted by the readout register and are then amplified, corrected and digitized to create the camera output. The readout register shifting is performed whilst the next row of pixels is being exposed. “Line rate” is the maximum rate at which readout and exposure can occur and is specified in kilohertz (kHz).
High line rates are required to “freeze” the motion of fast moving objects. Teledyne DALSA line scan cameras have line rates of up to 200 kHz, or 5 microseconds per rows of pixels imaged. Multiple “taps” are readout points along the readout register and these are used to boost the readout speed.
Line Scan Acquisition Interfaces
The camera’s pixel data are transmitted either to a frame grabber or a vision processor for processing. Cameras with three types of data transmission interfaces are provided by Teledyne DALSA.
It is possible to use Gigabit Ethernet (GigE) for data rates up to 80 Megabytes per second (MB/s). A standard GigE interface receives the camera data at the vision processor. This GigE interface is used on GEVA (GigE Vision Appliance) vision processors offered by Teledyne DALSA.
CameraLink interfaces receive the data through a frame grabber and have one of three formats. 255 MB/s is transmitted by the Base format using a single cable. The Medium format is 510 MB/s and the Full format transmits 680 MB/s, using two cables.
Pioneered by Teledyne DALSA, the HSLink interface transmits 6000 MB/s and needs a frame grabber to receive the data. For instance, the Piranha3 16k (16,384 pixels in a line) camera employs HSLink to acquire about 1,179,000,000 pixels per second at a 72 kHz line rate.
The transmitted lines of pixel values in the vision processor or frame grabber are accumulated into a frame – an image with the width (X dimension) of the line scan sensor and a user-specified height, which is often no more than 4,000 lines. As examples, frames are processed to carry out the machine vision task, reading text (OCR) or finding defects. Within in the vision processor, successive frames are overlapped for instance to detect a spot defect that covers the top edge of one frame and the bottom edge of the next one.
Time Delay and Integration
Time Delay and Integration (TDI) line scan cameras have 2 to 256 rows of pixels, arranged vertically. Photoelectrons from each pixel in each row are summed into the line of pixels “below”, in the direction of object motion. The summing is done by shifting the accumulated photoelectron charges.
Both summing and shifting are driven by the movement rate of the object, often signaled by pulses from a motion encoder. TDI cameras are capable of multiplying the exposure time by the number of rows of pixels and thus can provide high-contrast images even during a short exposure time. Another way to look at this is that rows of pixels are electronically moved in synchrony with the object’s movement being imaged and as a result the exposure time is increased without causing further blurring due to the movement of the object.
It must be noted that the exposure time of the camera is not dependent on the line rate, except that the exposure time needs to be less than the inverse of the line rate (seconds per line).
Color Line Scan Cameras
In color line scan cameras, there are rows of sensor pixels having different color filters or different color filters on pixels along the row, to detect light of different wavelengths. Generally, filters are perceived as red, green and blue (RGB), but certain applications, such as satellite remote sensing, use more and different types of filters.
Trilinear sensor pixels with Red, Green, Blue filters per line.
Bilinear sensor pixels with Green and alternating Red and Blue filters.
In particular, color line scan cameras are useful in imaging printed material. For this task, a “trilinear” camera with one filter type per line can be used, so that each image pixel has a complete set of RGB measures. The various lines of color sensor pixels are delayed in time, and hence they all view the same portion of the object being imaged.
For color line scan cameras, lenses should be color corrected so that they do not experience major chromatic aberration. Chromatic aberration takes place when a lens does not focus light of different wavelengths onto the same focal plane, and looks like color fringes around intensity edges.
If the camera is able to see an object at an angle off of the normal, that is perpendicular to the surface of the object, some perspective distortion might be there. Perspective distortion is in proportion to 1/distance to the object, and hence is somewhat different for the different color rows of pixels. This means, each row of pixels has a slightly different shape and size of pixel as projected on the surface of the object, and this can result in errors in high-accuracy color measurements. It must be ensured that the camera is perpendicular to the object’s surface – which is a good idea even if one is not using color.
Lighting for Line Scan Cameras
Light is required to view objects, and optics and lighting serve as “optical processors” to improve the interesting features of an object. Typically, line scan cameras use a “line light”, focused light from a series of LEDs, to light up the object along the line of object pixels being visualized. Line lights provide high intensity illumination required for short exposure times (fast camera line rates). Further, line lights can be “butted” together in order to give very long lights for viewing wide objects.
A high-powered line light from Metaphase Technologies, Inc.
Usually, line lights are positioned at a “high angle” above the object so “looking down” on the object, or at the back of the object being imaged to give “back lighting”. For instance, back lighting is used for detecting particle defects in transparent material or pinhole defects in opaque materials. For example, high angle illumination is used to obtain object outlines or colors or to detect surface defects.
The light intensity relies on the LEDs used, the efficiency of the lens, age and quality of the lamps and power flowing through the light. These variables make it hard to specify a light with just the right amount of intensity, and hence more illumination intensity than might be needed is often specified to be safe.
How to Select the Right Line Scan Camera for Your Application
Three factors should be considered when selecting a line scan camera such as sensitivity, size (in sensor pixels) and line rate.
“Sensitivity” asks, “Is the camera getting sufficient photons to execute the machine vision task?” It is usually difficult to answer this question only from component specifications, as there are too many variables. In practice, one estimates the light intensity required, specifies more light intensity than the estimate and then verifies that required sensitivity is obtained by testing. Some of the variables are as follows:
First, as stated before, the light intensity relies on a number of different variables. Second, the lens optics decreases the light to the camera. This reduction can be significant for high magnification lenses or when the lens is stopped down to obtain more depth of field. Third, “sensitivity” relies on the camera response– its potential to change photons to photoelectrons.
For better sensitivity, a camera can be used with larger sensor pixels. Line scan cameras have 100% “fill factor”, the percent of the sensor pixel that accumulates photons. Usually, area scan cameras have lower fill factors and smaller pixels, and make up for this loss in response by requiring longer exposure times. A TDI line scan camera can then be used to increase the exposure time by electronically moving the sensor lines of the camera with the object.
Size (in Sensor Pixels)
When determining the camera size (in sensor pixels), both the field of view (FOV) and minimum defect size need to be specified. The camera should have sufficient resolution to have at least three or four pixels “covering” the minimum defect size. If, for instance, the FOV is 12” and the minimum defect size is 0.005”, then:
(FOV/minimum defect size) x (3 pixels coverage)
(12/0.005) x 3 = 7200 pixels
One can use an 8K (8,192 pixel) line scan camera, or two 4K (4096 pixel) line scan cameras alongside with some overlap in their horizontal fields of view.
The FOV, object pixel size and part speed set the line rate. For instance, if the FOV is 12”, the part speed is 60” per second and an 8K (8,192 sensor pixels) camera is being used, then:
(Object pixel size in the FOV) = FOV/(camera size in sensor pixels)
= 12/8192 = 0.001465” object pixel size in the FOV
Line rate needed: 60”/0.001465” = 40.956 kHz
A Teledyne DALSA Piranha4 70 kHz camera can also be used.
Optics for Line Scan Cameras
The lens enhances sensitivity by collecting light from the object, and magnifies or minifies the FOV to match the size of the camera’s line sensor.
The magnification can be computed as:
Magnification = (camera pixel size in microns) / (FOV object pixel size)
By using the Piranha4 camera in the above example, this would be:
Magnification = (7.04 microns) / (37.211 = 0.001465 inches converted to microns) = 0.189
Therefore, a reduction in size of 1/0.188 = 5.286 from the “real world” FOV to the camera’s line sensor. This ratio (5.286 in this case) is the “inverse magnification”.
Next, one needs to specify a “working distance” – the distance from the camera’s face plate (the front, flat surface of the camera) – to the object being imaged (for instance, the web material). From this, the required focal length of the lens can be obtained:
Focal Length = (working distance) / (inverse magnification + 1)
In this case, a working distance of 314.28 mm = 12.37 inches is selected in order to give a 50 mm focal length – a standard focal length for lenses. In order to match a standard lens focal length, the working distance is often adjusted.
In addition to the focal length, it is essential to know the lens mounting type. Longer line scan sensors require an F-Mount, M72 or M42 mount lens, while smaller line scan cameras, up to 2,048 pixels in some cases, can use a C-mount lens with a 1” aperture.
Line Scan Camera Synchronization
It is important that line scan camera exposures are synchronized according to the movement of the object. This is often done with help the of an encoder that outputs a pulse for each specified amount of object motion. Following some number of encoder pulses, the line scan camera is finally triggered to take a line image.
“Square pixels” are usually preferred, that is pixels of equal height and width in the field of view. In order to obtain square pixels, the camera needs to be triggered each time the object moves a distance equal to the pixel size of the object. In the example given above, a pixel size in the field of view is 0.001465 inches which means for square pixels, the camera needs to be triggered each time the object moves that distance.
To set the line acquisition trigger distance, users can (1) choose or program the encoder’s pulses per revolution or distance, (2) choose the mechanical ratio between the object movement and encoder, and (3) program the number of encoder pulses between camera line acquisitions.
Based on the above example, assume that a rotary encoder is attached to a conveyer roller of known circumference and the object is on a conveyer belt. For square pixels, a camera trigger pulse is required every 0.001456 inches of conveyer belt movement. It is assumed that there is no mechanical slippage between the rotation of the encoder and the object movement.
An encoder with 4096 pulses per revolution was selected and the circumference of the conveyer roller is 3 inches (so diameter = circumference / pi = about 1 inch). The distance per encoder pulse is:
Distance per pulse = (roller inches per revolution) / (encoder pulses per revolution)
= 3 inches / 4096 pulses = 0.0007324 inches per pulse
This is too small of a distance by about a factor of 2. By fixing the roller circumference, the encoder inputs or a PLC’s “axis” inputs can be used on a Teledyne DALSA Xcelera frame grabber to “divide” (actually, count-down) the encoder pulses by a factor of 2. Alternatively, an encoder with 2048 pulses per revolution can be used. Either would give:
Distance per pulse = 2 * 0.007324 = 0.0014648 inches
Pixel Correction with Line Scan Cameras
Hardware correction of individual pixels’ responses is an advantage of a line scan camera. Many sources of variation are there in pixel responses. Lighting is never fully uniform –some variation is usually seen because of variations in brightness of individual lamps in the light and some reduce in intensity towards the edges of the light.
Most lenses attenuate light away from the optical axis (the center line through the lens). Individual pixels in the camera have some variation in their offset and gain due to variations in the camera manufacturing process.
Each offset and gain of the pixel is modified to compensate for these sources of variation. This can be realized by exposing all pixels to the highest (but not saturating) intensity on the object and the lowest intensity on the object. The lowest intensity value is subtracted by offset correction. Gain correction sets the range between the highest intensity value and the (corrected) lowest intensity value to be constant across sensor pixels.
Pixel response before gain and offset correction.
Pixel response after gain and offset correction.
Line Scan Cameras at Work
Two “real world” applications of line scan cameras are described here.
Multiple Camera Web Inspection
Line scan technology is the ideal method in continuous web applications. As an example, a production line for plastic film was inspected; the film measures 12 feet wide and moves at 350 linear feet per minute. The objective was to look for small defects, such as dirt, holes or contaminants. However, attempting this level of inspection with an area scan camera is quite difficult because of the high speed and wide area of the web. This cannot be done by human vision inspection.
Defects are categorized based on contrast type and size range, such as from 50 to 100 microns or 200 microns. A dark object might be a contaminant and a bright object might be a hole. This level of detail and web width requires six Gigabit Ethernet 2048-pixel line scan cameras. Since backlight is used, a line light shines through the web and into the camera:
The aggregate data rate is 73 million pixels per second. An image for processing is 3560 lines.
With a continuous web application, it is not possible to reject a defect instantly; the web is moving too fast to halt the process. Instead, the position and type of defects are recorded and described by a “roll map”. This screen shot given below reveals the portion of the roll map made by camera 5:
This demonstrates a real-time view (on the left). The roll map, on the right, the x axis shows the defect position and the y axis identifies linear footage (in roller revolutions). When the product roll is sent for finishing, defects can be identified as the product is unwound.
Color Line Scan
Medical trays incorporating grommets, rivets and brackets of different colors are inspected using a color line scan camera. These elements cannot be easily differentiated and studied using gray-scale (intensity) images.
Operator’s view of the medical tray; a missing part indicated by a thick, red box. (upper, right)
The manufacturer creates trays of many different sizes, the largest of which is 18 inches in width and 27 inches in length. To accommodate the largest medical tray, a 2K color line scan camera is mapped to 20 inches on the horizontal. With regards to line scan technology, there is no limit on the vertical size. Two 24 inch line lights light up the whole tray, which is shifted along a conveyor. Then, an encoder on the conveyor is synchronized, so that the color line scan camera triggers as the tray comes into the field of view.
Line scan technology is suited for applications that image large objects, are high-speed, are high-resolution, require perfect images or require real-time correction of pixels’ responses. The camera features a single line of pixels, and either the object or camera is moved perpendicular to that line of pixels to create a two-dimensional image of the object
In order to achieve high brightness, “line lights” are focused where the line scan camera is visualizing the object. Longer exposure times are given by time delay and integration (TDI) cameras by electronically moving lines of pixels in synchrony with the object’s movement.
Applications where line scan cameras have been shown to be the best choice include inspecting web materials, as in paper or plastic manufacturing, analyzing continuous “objects” such as rail and road inspection, or where high-quality images are needed as in printing inspection.
When using line scan technology, care must be taken to synchronize the camera acquisition to the object's movement. Starting costs may be higher than for frame cameras, but line scan technology solves problems that frame cameras cannot easily solve or solve at all.
This information has been sourced, reviewed and adapted from materials provided by Teledyne DALSA.
For more information on this source, please visit Teledyne DALSA.