• How can I run a line scan camera at near to zero Hz (zero lines per second)?

For some cameras, running the camera at a line rate near zero is not a problem. For example:

1.	L304k, L304kc, L400k, and L800k cameras have no minimum required line rate when an external trigger (ExSync) signal is used to trigger line acquisition. Keep in mind that for proper operation, the exposure time should be at least 10% of the line period. And if these cameras are used in free run, there is a 10 Hz minimum line rate.
2.	Runner cameras have no minimum line rate when an external line start trigger (ExLineStTrig) signal is used to trigger line acquisition. However if an external line start trigger is not used, there is a minimum 100 Hz line rate.

On some cameras, there is an absolute minimum line rate. For example:

1.	L100k, L301k, L301kc and sprint cameras all have a minimum line rate of 1KHz.

So why do some cameras have an absolute minimum line rate?

Basic Camera Principles

The principle of how a camera works is that during line exposure, photons from a light source strike the pixels in the camera's sensor and generate electrons. At the end of each line exposure, the electrons collected by each pixel are transported to an analog-to-digital converter. For each pixel, the converter provides a digital output signal that is proportional to the number of electrons collected by the pixel.

Below Minimum Line rates

If a camera is triggered at a rate below the specified minimum, it is much easier to fall into an over exposure situation. This happens due to an effect called "shutter inefficiency". The electronic shutter on digital cameras is not 100% efficient, and the pixels in the camera will collect some photons even when the shutter is closed. At very low line rates, you have long periods of time between exposures when the shutter is closed but the pixels are still collecting some photons and generating electrons. When the electrons collected with the shutter closed are added to the electrons collected during an exposure, the electrons can flood the electronics around the pixel.

After an Over Exposure

After an over exposure or with a trigger rate below 1kHz, it takes several readout cycles to remove all the electrons from the pixels and the electronics. For this reason, gray values will be abnormally high during the first several readouts after an over exposure.

• Can you tell me how to close the control channel between a Basler GigE Vision camera and an application?

• How can I interface a Basler GigE Vision camera with my Stemmer Common Vision Blox 9.0.2 Software?

Application notes are available that provide a detailed description of how to interface Basler GigE cameras with Stemmer Common Vision Blox 9.0.2 Software.

• How can I use pylon's DirectShow filter to capture video from my Basler Camera

Application notes are available that describe using pylon's DirectShow filter along with some common open source software to capture video from Basler cameras.

• Can you tell me which network adapters are compatable with the Basler pylon performance driver?

Basler provides two GigE Vision network drivers for interfacing Basler GigE Vision cameras:

1.	The Basler filter driver is a basic GigE Vision network driver that is compatible with all network adapters. The advantage of the filter driver is its extensive compatibility.

2.

The Basler performance driver is a hardware specific GigE Vision network driver. The performance driver is only compatible with network adapters that use specific Intel Pro 1000 chipsets ("compatible chipsets"). The advantage of the performance driver is that it significantly lowers the CPU load needed to service the network traffic between the PC and the camera(s). It also has a more robust packet resend mechanism.

To take advantage of the benefits of the Basler performance driver, we recommend using Intel Pro 1000 network adapters. These adapters generally work well with the performance driver. However, since the Intel Pro 1000 series has changed over the time, it may happen that the Basler performance driver does not support your particular Intel Pro 1000 adapter.

To make sure that the Pro 1000 network adapter you are using is compatible, consult the table below that lists the currently supported Intel Pro 1000 chipsets and their corresponding Hardware IDs. Note that some chipsets are compatible with pylon version 2.1, but not with version 2.0.

In the following table:

1.	Yes = is compatible
2.	Yes/M = is compatible but requires manual installation
3.	No = is not compable

		Intel Pro 1000 Chipset	Hardware ID	Pylon  2.0	Pylon  2.1
		82540EM	PCI\VEN_8086&DEV_100E	Yes	Yes
		82540EP_EL	PCI\VEN_8086&DEV_101E	Yes	Yes
		82541GI	PCI\VEN_8086&DEV_1076	Yes	Yes
		82541GI_LF	PCI\VEN_8086&DEV_107C	Yes	Yes
		82545EM	PCI\VEN_8086&DEV_100F	Yes	Yes
		82545GM	PCI\VEN_8086&DEV_1026	Yes	Yes
		82563EB/80003ES2	PCI\VEN_8086&DEV_1096	No	Yes
		82567/ICH9_IGP_AMT*	PCI\VEN_8086&DEV_10BE	No	Yes/M
		82567/ICH9_IGP_AMT*	PCI\VEN_8086&DEV_10F5	No	Yes/M
		82571EB	PCI\VEN_8086&DEV_105E	Yes	Yes
		4-port<br>(2 X 82571EB)	PCI\VEN_8086&DEV_10A4	Yes	Yes
		4-port LP<br>(2 x 82571EB)	PCI\VEN_8086&DEV_10BC	Yes/M	Yes
		82572EI	PCI\VEN_8086&DEV_10B9	Yes	Yes
		82572EI-Copper	PCI\VEN_8086&DEV_107D	Yes	Yes
		82573E	PCI\VEN_8086&DEV_108B	Yes	Yes
		82573E-IAMT	PCI\VEN_8086&DEV_108C	Yes	Yes
		82573L	PCI\VEN_8086&DEV_109A	Yes	Yes
		82574L	PCI\VEN_8086&DEV_10D3	No	Yes

* pylon operation with this chipset is unreliable


To check the Hardware IDs for your network adapter:

1. Click Start > Run.

2. Type in: devmgmt.msc

3. Click the OK button. The device manager will start.

3. Expand the node for Network Adapters.

4. Right click on the name of your Intel Pro 1000 adapter
and select Properties from the drop down menu.

5. Click the Details tab and make sure that Hardware IDs is selected in the drop down list.

On PCs with a Windows™ OS, if you configure a Basler GigE camera for a persistent (fixed) IP address and the address is in the range normally reserved for "multicast" IP addresses (224.0.0.0 to 239.255.255.255), the camera will not be discoverable by pylon, even when you use the pylon IP Configuration Tool. This situation occurs because Windows rejects all incoming IP packets from any device (such as a GigE camera) with an IP address in the multicast range. You can find some good basic information about IP multicast on Wikipedia.

Click here to open the Wikipedia article.

The pylon IP Configuration Tool works by sending UDP broadcast messages to all attached cameras and waiting for the cameras to answer. But since Windows rejects packets from devices with IP addresses in the multicast range, answers from any camera with an IP address in the range will never reach the configuration tool.

With the current package of pylon tools, there is no way to discover a camera with an IP address set in the multicast range. However, the Basler pylon API does provide a method for accessing a camera by its MAC address and for forcing a change to the camera's IP address (the FORCEIP_CMD). This will let you set the camera back to a state where it is discoverable.

A programming sample is available that illustrates how to use the pylon C++ API to set the camera's IP configuration. It also illustrates how to use the Force IP command. The programming sample is based on the Basler pylon 2.0 C++ SDK and to build the entire sample project, you must have the pylon 2.0 SDK installed on your PC. The sample also includes a prebuilt "Simple IP Configuration Tool" executable which will run on PCs that have either the pylon SDK or Basler's free pylon 2.0 runtime package installed.

You can use the link below to download the programming sample:

• Can you give me some advice about using the frame transmission delay and interpacket delay parameters on my Basler GigE Cameras to control frame transmission?

Application notes are available that give a good overview of what these parameters do and provide advice about how you can best use the parameters in some specific situations.

Some users of Basler's Pylon 2.0 software may want to build their own applications without the need to buy Microsoft Visual C++ Studio. Application notes are available that describe a way to build pylon based applications for free using Microsoft Visual C++ Express and the Microsoft Platform SDK.

Click here to download the application notes.

• How can I interface a Basler GigE Vision camera with my National Instruments Vision 8.2.1 Acquisition Software?

Application notes are available that provide a detailed description of how to interface Basler GigE cameras with Vision 8.2.1 Acquisition Software from National Instruments.

Some detailed information about GigE Vision can be found in the White Paper downloads section of our website. Information about CPU load and latency can also be found there. Click here to go to the white paper downloads.

The table below presents an overview of the most important differences between the two interfaces:

		GigEVision	IEEE 1394b
	Cable length	Up to 100 meters per cable. Multiple cables can be connected using switches or repeaters, thus allowing a virtually unlimited cable length. Standard CAT6 Ethernet cable can be used.	Up to 4.5 meters per cable. With special cables and a repeater, the max. cable length can be extended up too 14.5 meters. Standard FireWire cables can be used.
	Power supply	Camera power must be supplied via a power connector on the camera. An additional cable for the power supply is required.	Camera power is supplied over the 1394 cable
	Broadcasting commands(one application sends the same command, e.g., a software trigger, to all cameras connected to one network/bus)	Currently not supported, but in preparation	Fully supported
	Max. available bandwidth	Up to 1 gigabit per sec	1394a - up to 400 megabits per sec 1394b - up to 800 megabits per sec
	CPU load(the CPU load caused by streaming data from the camera to the PC is highly dependent on the PC hardware and camera settings)	Low (5-10% is typical)	Even lower (> 1%)
	Latency when sending a command from the PC to the camera(e.g., a software trigger)	600-900 µs (jitter of about 30%)	350-500 µs (jitter of about 30%)
	Latency when sending image data from the camera to the PC(when image will be present in PC memory is highly dependent on the PC hardware and camera settings)	13-14 ms	13-14 ms
	Accessory hardware	All required hardware is available as standard, industrially proven devices: 1. Gigabit Ethernet network interface connector (also called the Ethernet card, NIC, or network adapter) is most likely already part of the PC 2. Switch (if needed to connect multiple cameras to one NIC) 3. Standard CAT6 Ethernet cable	The required hardware is typically derived from the consumer market: 1. FireWire adapter card is required 2. Standard FireWire cables 3. Hub (if needed to connect multiple cameras to one FireWire card or to extend the max. cable length) 4. Repeater (if needed to extend the max. cable length)

• Can you tell me which compiler works with which Basler Camera Driver/API?

Basler provides two different high level C++ APIs for interfacing with our IEEE 1394 (FireWire) and GigE Vision cameras: the classic BCAM 1394 API and the GenICam compliant pylon API.

The table below shows which API supports which camera interface and uses which compiler.

		BCAM 1.8	BCAM 1.9	pylon 1.0	pylon 2.0	pylon 2.1
	Supported Camera Interfaces	1394a	1394a 1394b*	1394a 1394b* GigEVision	1394a 1394b** GigEVision	1394a 1394b** GigEVision
	Visual C++ 6.0 (part of Visual Studio 6.0)	Y	Y	N	N	N
	Visual C++ 7.0 (part of Visual Studio 2002.NET)	Y	Y	N	N	N
	Visual C++ 7.1 (part of Visual Studio 2003.NET)	Y	Y	Y	Y	Y
	Visual C++ 8.0 (part of Visual Studio 2005)	Y***	Y***	Y	Y	Y
	Visual C++ 9.0 (part of Visual Studio 2008)	Y***	Y***	Y****	Y****	Y

• Can you clarify some of the buzzwords about DirectX, DirectShow, and WDM?

DirectX

Microsoft DirectX represents a collection of technologies established by Microsoft. It provides accelerated direct and uniform access to the video and audio hardware installed on a system. Since the introduction of Windows 98, DirectX has been an integral component of all Microsoft operating systems. You can view more detailed, official information about DirectX on the Microsoft website at: http://msdn2.microsoft.com/en-us/library/ms786508(VS.85).aspx

DirectShow

DirectShow is the generic term for those parts of the DirectX API that control the behaviour of video and audio data streams. Since the release of Windows 98, DirectShow has been established as a standard for multimedia oriented image processing. DirectShow enables any DirectShow compliant device (e.g., a camera) to operate with any DirectShow compliant software (e.g., DirectShow based image processing libraries such as MontiVision. See: http://www.montivision.com).

WDM

WDM is the abbreviation for the Windows Driver Model. The WDM describes a model driver architecture for Windows 2000 and its successors. A WDM streaming driver controls the processing and transport of streaming data at the operating system level ("kernel mode") instead of at the application level ("user mode").

• I’m using a Fast Ethernet switch or hub to connect my camera to a GigE network adapter, but my camera isn’t discovered. Why not?

• I can’t get the full frame rate on my GigE camera. Why not?

• Why am I having problems using multiple network adapters in my PC?

• I’ve installed the Basler pylon Viewer and a pylon network driver (either the Basler Filter Driver or the Basler Performance Driver), but I do not see my camera in the "view tree" of the pylon Viewer. Why not?

• Why can’t I capture images with the pylon Viewer?

• How can I run my scout 1394b camera at full FireWire-b speed?

• Can I use the BCAM 1394 Driver to directly write to or read from the registers in my Basler 1394 camera?

• On my Basler 1394 camera, how can I save my current camera settings or configure the camera to load my preferred settings at power up?

The set of parameters that a Basler camera uses for camera control is known as the "work set" of parameters. When you use the Basler BCAM 1394 Driver or another tool to change parameter values, you are changing the settings in the work set. The work set resides in the camera's volatile memory and is lost at power off or camera reset.

Most Basler 1394 cameras have a feature that lets you save the current work set into one of three "memory channels". A set of parameters saved in one of the memory channels is known a "user set". The memory channels are part of the camera’s internal, non-volatile memory and the user sets stored in the memory channels will be retained when the camera is powered off or reset. No memory channel data will be stored on your PC.

Each camera also has a set of factory parameter settings stored in a separate memory channel. The memory channel containing the "factory set" can't be overwritten.

You can save up to three different user sets in the camera's memory channels. If you desire, you can load one of the saved user sets from a memory channel into the camera's work set. And if you want to restore the camera to its initial factory settings, you can simply load the camera's factory set into the work set.

• How can I build BCAM 1394 applications using Basler's SDK and Visual Studio 2005.NET?

• How can I avoid video or data corruption on notebook computers using an IEEE 1394 connection?

Avoiding Data or Video Corruption

Corrupted data or video when using a notebook's IEEE 1394 port can be caused by too much latency in the C3 power state transition, which results in buffer under runs. In other words, the interrupts associated with the processor's ability to dynamically change speeds conflict with the high demand on the processor that occurs when streaming video or data across the IEEE 1394 port. When operating a Basler 1394 camera at high data rates, this behavior can be observed, for instance, as a jittering test image when using the BCAM viewer. We found this behavior with several different notebook computers including Dell, Toshiba, Acer, Gericom, and others. The problem may disappear or may become less noticeable when a USB device such as a mouse or a memory stick is connected to the notebook PC.

A better workaround, however, is to change the C3 latency setting in the Windows® Registry. You can either change the setting manually or you can download and run a file that will make the change automatically.

Caution! The following procedures contain information about editing the Windows registry. Basler does not guarantee success or support these actions. Any use of the information provided herein is performed at your own risk. You should make a backup copy of the registry files before executing any of the following steps. Incorrect use of the registry editor and editing the registry files can cause serious problems that may require a complete reinstall of your operating system. Basler assumes no responsibility, expressed or implied, regarding the consequences of any action taken as a result of the information provided herein.

Changing the Setting Manually

To change the C3 latency setting in the Windows registry, perform the following steps:

1. Click the Start button and then click Run.

2. In the Open box, type: regedit

3. Click OK. The registry editor will open.

4. In the registry editor, locate the following key:

     HKEY_LOCAL_MACHINE
SYSTEM
CurrentControlSet
Control
Processor

5. In the Processor key, right-click CStateFlags and select Modify.

6. Change the Value data to 8.

7. Exit the Registry Editor.


Note: The Processor key may not exist. If it does not exist, you will need to create it. To create the key, perform the following steps:

1. Right-click the Contol key and select New and Key from the menu.

2. Rename the new key as: Processor.

3. Click on the Processor key. From the Edit menu, select New and DWORD value.

4. Rename the new DWORD value to: CStateFlags.

5. Right-click CStateFlags and select Modify.

6. Change the Value data to 8.

7. Exit the Registry Editor.

After you make any changes to the registry, you must restart the computer for the changes to take effect.


Making the Changes by Running a File

As an alternative to manually changing the registry, you can download the file below and run it to make the changes. If the Processor key is not present in the registry, running this file will create the key and will set its value. If the Processor key already exists, running this file will simply change the value of the existing key.

• Do I need an extra power supply when I'm using a Basler IEEE 1394 camera with a laptop?

• I see a "moiré" pattern in my captured images. What can I do?

• I'm having some problems after installing the BCAM driver. What can I do?

BCAM Installation Problems

After a new installation of the BCAM driver or after updating an existing BCAM driver installation to the current version, users sometimes encounter problems. These instructions will help you solve the most typical problems. The instructions are written for Windows XP, but if you are operating a system with Windows 2000, the solutions work in a very similar manner.

If you are using a laptop:

If you are using a laptop, please make sure that the camera is being properly supplied with power. Many IEEE 1394 PCMCIA cards only have a 4 pin connector instead of the normal 6 pin connector. The two missing pins are for the wires used to supply power to the camera. If your PCMCIA card has only 4 pins, it will not supply power to the camera and your camera will not work. In this case, you can either switch to a card that does supply power to the camera or you can add a powered IEEE 1394 hub to your system.


Common Error Messages:

After installing the driver or an update, you may see one of these error message when you open the BCAM viewer:

"The Versions of the BCAM API(1.6) and BCAM Driver(1.8) don't match"

"BCAM is not compatible with driver 0.0"

"Die Netzwerkanforderung wird nicht unterstuetzt"

All of these messages have the same origin, i.e., the camera driver is not properly associated with your camera. To correct this error, you must change the driver association.

To begin the process, click Start > Control Panel > System > Hardware > Device Manager. A Device manager window will open.

• How can I save image data deeper than 8 bits as a TIFF?

Saving Image Data Deeper Than 8 Bits as a TIFF Using Libtiff with Visual Studio 7

This FAQ describes how to save 16 bit monochrome image data to a conforming TIFF with libtiff. Once the image is saved, you can do further image post processing with other applications like ImageJ. Libtiff is open source software, but is not GPL.

Integrating Libtiff into your Solution

First, you must get libtiff. I suggest that you download Version 3.7.2 of the original sources from: ftp.remotesensing.org/pub/libtiff/old/. When your download is complete, copy the libtiff sources (\tiff-3.7.2\libtiff) into your solution directory (\libtiff). There are two configuration files with pre-processor definitions: tif_config.h and tiffconf.h. Since we are working with Visual Studio (on an Intel/AMD platform I assume), rename:

	tiffconf.h		to   tiffconf.h.old
	tiffconf.h.vc		to   tiffconf.h
	tif_config.h.vc		to   tif_config.h

Next, set up a Win32 project named libtiff with a static library and no precompiled header.
NOTE: It is not absolutely necessary for you follow this part of the FAQ in all of its particulars. If you need to vary the procedure (e.g., by building a dynamic link library instead of a static library), then do so. You can also use the attached make-file to create a library. That’s just the way I made it and it worked.

Now add all the libtiff-sources to your library, except for the dispensable OS-specific ones:
tif_acorn.c, tif_apple.c, tif_atari.c, tif_msdos.c, tif_stream.cxx, tif_unix.c, tif_win3.c

To access libtiff support in your code, add libtiff-path to include-paths, include <tiffio.h>, and link your application against the created libtiff library.

The Sample Code: Save/load Image Data to/from your HDD

I’ve prepared some sample code to illustrate working with 16 bit images. Click the link below to download the sample.

A TIFF (Tagged Image File Format) file consists of anterior tags that describe the following image data whether it is compressed or it is raw (this FAQ only deals with raw image data). So the method to save image data needs to know the destination filename, a pointer to the image data, the dimension, and the data depth that are assigned to the image data.

bool save16mono ( const char * szFilename,

			const unsigned short * pImageBuf,
			const unsigned int width,
			const unsigned int height,
			const unsigned int bppReal );

The method to load the image data only needs to know which file to load to which memory area and tells you in which dimension and at which data depth the image was stored.

bool load16mono ( const char * szFilename,

			unsigned short * pImageBuf,
			size_t SizeImageBuf, // for buffer overrun prevention
			unsigned int * pWidth,
			unsigned int * pHeight,
			unsigned int * pBppReal );

The TIFFSetField()/TIFFGetField() routines give you the ability to write/read such a tag (e.g. TIFFTAG_IMAGEWIDTH-tag for image’s width).
The TIFFWriteScanline()/TIFFReadScanline() routines provide writing/reading an image line to/from a TIFF file.
Both of these routines need a TIFF* file handle, which TIFFOpen() delivers. The semantic is just like the one for fopen()/fclose().

My sample code first generates a 10 bit deep image gradient (400 x 300 pixels, pExampleBuf) and then saves the image data, loads the image into another buffer (pLoadBuf), and compares the generated and loaded buffers to each other. In some cases, the image data you want to save/load is not necessarily 16 bits deep (it could be anywhere between 8 bits and 16 bits). So, for example, if you have 10 bit image data for one pixel:

If you save the data as shown above, most image processing applications will clip the image data to 8 bits in order to display them. So what you’ll see are the upper 8 bits:

That includes nearly nothing from the original picture. To avoid this problem, we must shift the original image data:

• What is the RGB color space?

RGB Color Space

Because the human eye only has color sensitive receptors for red, green and blue, it is theoretically possible to decompose every visible color into combinations of these three “primary colors.” Color monitors, for instance, can display millions of colors simply by mixing different intensities of red, green and blue. It is most common to place the range of intensity for each color on a scale from 0 to 255 (one byte). The range of intensity is also known as the “color depth.”

The possibilities for mixing the three primary colors together can be represented as a three dimensional coordinate plane with the values for R (red), G (green) and B (blue) on each axis. This coordinate plane yields a cube called the RGB color space:

If all three color channels have a value of zero, it means that no light is emitted and the resulting color is black (on a monitor, for example, it cannot be blacker than the surface of the monitor producing 0 light). If all three color channels are set to their maximum values (255 at a one byte color depth), the resulting color is white. This type of color mixing is also called “additive color mixing”:

• How does the YUV color coding work?

YUV Color Coding

A CCD or a CMOS sensor alone is not able to detect the color of incident light. In reality, each pixel in the sensor simply detects the intensity of the incident light. But when a color pattern filter is applied to the sensor, each pixel becomes sensitive to only one color - red, green or blue. The following table shows the color arrangement of a “Bayer Pattern” filter on a sensor with a size of X x Y (with X and Y being multiples of 2).

Since the arrangement of the colors in the Bayer pattern filter is known, a PC can use the raw information transmitted for the pixels to interpolate full RGB color information for each pixel in the sensor. Instead of using the raw sensor information, however, it is more common to use a color coding known as YUV. The block diagram below illustrates the process of conversion inside a Basler IEEE 1394 camera. To keep things simple, we assume that the sensor collects pixel data at an 8 bit depth.

As a first step, an algorithm calculates the RGB values for each pixel. This means, for example, that even if a pixel is sensitive to green light only, the camera gets full RGB information for the pixel by interpolating the brightness information from adjacent red and blue pixels. This is, of course, just an approximation of the real world. There are many algorithms for doing RGB interpretation and the complexity and calculation time of each algorithm will determine the quality of the approximation. Basler IEEE 1394 color cameras have an effective built-in algorithm for this RGB conversion.

A disadvantage of RGB conversion is that the amount of data for each pixel is inflated. If a single pixel normally has a depth of 8 bits, after conversion it will have a depth of 8 bits per color (red, green and blue) and will thus have a total depth of 24 bits.

YUV coding converts the RGB signal to an intensity component (Y) that ranges from black to white plus two other components (U and V) which code the color. The conversion from RGB to YUV is linear, occurs without loss of information and does not depend on a particular piece of hardware such as the camera. The standard equations for accomplishing the conversion from RGB to YUV are:

Y = 0.299 R + 0.587 G + 0.114 B
U = 0.493 * (B - Y)
V = 0.877 * (R - Y)

In practice, the coefficients in the equations may deviate a bit due to the dynamics of the sensor used in a particular camera. If you want to know how the RGB to YUV conversion is accomplished in a particular Basler camera, please refer to the camera’s user manual for the correct coefficients. This information is particularly useful if you want to convert the output from a Basler IEEE 1394 camera from YUV back to RGB.

The diagram below illustrates how color can be coded with the U and V components and how the Y component codes the intensity of the signal.

This type of conversion is also known as YUV 4:4:4 sampling. With YUV 4:4:4, each pixel gets brightness and color information and the “4:4:4” indicates the proportion of the Y, U and V components in the signal.

To reduce the average amount of data transmitted per pixel from 24 bits to 16 bits, it is more common to include the color information for only every other pixel. This type of sampling is also known as YUV 4:2:2 sampling. Since the human eye is much more sensitive to intensity than it is to color, this reduction is almost invisible even though the conversion represents a real loss of information. As defined in the DCAM specification, YUV 4:2:2 digital output from a Basler camera has a depth that alternates between 24 bits per pixel and 8 bits per pixel (for an average bit depth of 16 bits per pixel).

As shown in the table below, when a Basler camera is set for YUV 4:2:2 output, each quadlet of image data transmitted by the camera will contain data for two pixels. In the table, K represents the number of a pixel in a frame and one row in the table represents a quadlet of data transmitted by the camera.

• Basler cameras run hot. Is this really a problem?

• What is FireWire?

• Does CMOS technology have advantages over CCD technology?

• How does 3-Chip color compare with 1-Chip color and can you discuss some of the advantages and disadvantages of each?

3 Chip vs. 1 Chip Color

Three chip color cameras always contain a prism which divides the incoming light rays into their red, green and blue components. Each chip then receives a single color at full resolution.

One chip area scan cameras use a single sensor that is covered by a color filter with a fixed, repetitive pattern. Filters with several different patterns are used but the Bayer color filter is the most common. The illustration to the right shows a portion of the Bayer filter. When a color filter is used with a single sensor, each individual cell in the sensor gathers light of only one particular color. To reconstruct a complete color image, an interpolation is needed. The red, green and blue information is interpolated across several adjacent cells to determine the total color content of each individual cell.

One chip line scan cameras use a sensor that has three rows of cells, a red rwo, a green row and a blue row. As an area on an object moves past the camera, the area is examined first by the cells in the red row, second by the cells in the green row and third by the cells in the blue row. The information from the red, green and blue cells is then combined to produce a full color image.

3-Chip Color Advantages:

1.	Full resolution RGB Images
2.	Easier software handling of the data output

3-Chip Color Disadvantages:

1.	High camera cost due to the need for a prism and three sensor chips
2.	Large camera housing needed for prism and sensors
3.	Typically require expensive, special optics
4.	High weight

1-Chip Color Advantages:

1.	Much less expensive
2.	Smaller size
3.	Lower weight

1-Chip Color Disadvantages:

1.	For area scan cameras, an interpolation algorithm must be run to achieve full color resolution
2.	For line scan cameras, spatial correction must be done to combine the color data from the three sensor rows

• What is sensitivity and why are sensitivity statements often misleading?

Sensitivity

The response curve for a light sensitive sensor can be divided into three parts: the dark area, the linear area and the saturation area. A typical response curve is shown in the graph below.

The dark area of the response curve shows the sensor's response to very low light. The output of the sensor in the dark area is very low, is noisy and is unpredictable. As you gradually increase the light falling on a sensor, you will find a point where the output of the sensor begins to increase predictably as the amount of light increases. This point is called the Noise Equivalent Exposure (NEE).

After the NEE point is reached, the output of the sensor becomes linear. The output remains linear until a point called the Saturation Equivalent Exposure (SEE) is reached. At this point, increasing the light intensity results in a nonlinear increase in the sensor output.

The gradient of the linear portion of the sensor's response curve is commonly referred to as sensitivity and is usually measured in V/µJ/cm2. The higher a sensor's output voltage is for a given amount of light, the higher its sensitivity.

But when you are discussing sensors, talking about sensitivity alone does not make sense. For one thing, NEE is also very important. Since a sensor with a high NEE will be blind at low light levels, NEE should be as low as possible.
Another point to consider is that a digital camera is a system and that sensor sensitivity is just one of the factors involved in the output signal from the camera. Electronic devices in the camera such as Analog to Digital converters and amplifiers also influence the output signal. At Basler, we feel that a camera's "responsivity" is a better measure of camera performance. We also think that since our cameras are digital, responsivity should be stated as DN/µJ/cm2 (DN stands for digital number). The graph below shows a responsivity curve.

• What is a Signal-to-Noise Ratio and how can I improve it?

• What is PRNU?

• What is the Area of Interest (AOI) feature?

• What is binning?

Binning in CCD Cameras

Binning increases the camera's sensitivity to light by summing the charges from adjacent pixels in the CCD sensor into one pixel. There are three types of binning available: horizontal binning, vertical binning, and full binning.

With horizontal binning , pairs of adjacent pixels in each line of the sensor are summed (see the drawings below). With vertical binning, pairs of adjacent pixels from two lines in the sensor are summed. Full binning is a combination of horizontal and vertical binning in which four adjacent pixels are summed.

• What are color filters?

Color Filters for Single-Sensor Color Cameras

In general single-sensor color cameras use a monochrome sensor with a color filter pattern. Another way to achieve a color image with only one sensor would be to use a revolving filter wheel in front of a monochrome sensor, but this method has its limitations.

With the color filter pattern method of color imaging, no object point is projected on more than one sensor pixel, that is, only one measurement (for a single color or sum of a set of colors) can be made for each object point.

There are several different filter methods for generating a color image from a monochrome sensor. In the following some frequently used filter arrangements are detailed.

Bayer Color Filter (Primary Color Mosaic Filter)

The following table 1 shows the filter pattern for a sensor of size xs x ys (xs and ys being multiples of 2):

Complementary Color Mosaic Filter

The following table 2 shows the filter pattern for a sensor of size xs x ys (xs and ys being multiples of 2):

This is basically the same arrangement as the Bayer filter pattern, but instead of using primary colors (R, G, B) it works with complementary colors (magenta, cyan, yellow). The reason for this is that a primary color filter blocks of 2/3 of the spectrum (i.e. green and blue for a red filter) while a complementary filter blocks only 1/3 of the spectrum (i.e. blue for a yellow filter). Thus, the sensor is 2 times more sensitive. The tradeoff is a somewhat more complicated computation of the R, G, B values, requiring the input of each complementary color.

Primary Color Vertical Stripe Filter

Table 3 shows the filter pattern for a sensor of size xs x ys (xs being a multiple of 4):