Phenomenon of interference which occurs when a signal being sampled contains frequencies that are higher than half the sampling frequency. Typically can be seen as ragged edges on horizontal lines.
If an excessive light is behing the center object, it is necesary to prevent the center object from becoming too dark. The cameras equipped with automatic back light compensation usually increase contrast locally in the center of the field of view to help avoid silhouetting.
A special operating mode of the electronic shutter of a CCD camera. The shutter timing is automatically adjusted to maintain, as far as possible, the same video signal level from the camera irrespective of scene illumination. Allows the use of a fixed iris lens under variable lighting conditions.
CCD sensor size, field of view, working distance, resolution and lens magnification are closely related characteristics of the camera. By using a smaller size CCD sensor, the field of view of the camera is decreased when using same lens. With smaller size CCD sensor, smaller details may be resolved from the same working distance.
Conventional CCD sensors (in opposite to progressive scan ones) are designed for use in interlaced scan systems of video and TV. These systems have two fields which are interlaced to form a full frame image (see Scanning). This technique brings some problems and image quality limits, which restrict the usability of such chip in image processing tasks, where high resolution scanning of moving objects is needed. Then one have to decide whether to emphasize on high vertical resolution or on low motion blur and choose one of the charge accumulation modes.
A device which converts light into electrical energy in a CCD camera. A CCD consists of a two-dimensional matrix of many thousands of individual photosensitive elements. The camera optics focus the scene onto the matrix and each element generates a charge which varies with the intensity of the light it receives. These charges are passed out, one by one, row by row, from a single connection to form a continuous analogue signal. This charge/discharge process is continuously repeated, normally at field rate (see also Progressive scan).
The color cameras use similar CCD chips as the monochrome cameras. There are various methods how to acquire the color information:
The transmission of color TV signal had to be solved in a way compatible with existing monochrome systems. That's why the color information is decomposed to its luminance Y and two color carrying signals R-Y and B-Y. All R, G and B image components in RGB color model. Both color R-Y and B-Y components are modulated to the luminance Y signal. The color carrier frequency (used for demodulating) is read out from short color synchronizing impulse, «burst», presented during horizontal retrace. That's why color signal can be watched in monochrome TV (as b/w image, of course) and vice versa.
The purpose of a color model (also color space) is to facilitate the specificaion of colors in some standard, generally accepted way. In essence, a color model is specification of a 3-D coordinate system and a subspace within that system where each color is represented by a single point.
There are several types of connectors possessed by most cameras and framegrabbers.
Direct Memory Access. A method by which data can be transferred from a device (e.g. framegrabber) to the computer memory and vice versa without processor intervention. As a result, the processor is free for other tasks and the data transfer is faster.
Donpisha means «immediate» in Japanese. A Donpisha shutter can be triggered to operate at a particular point in time and is used to shoot moving objects without a time delay. A fixed-position camera fitted with a Donpisha shutter can capture flicker-free images of fast moving objects.
The major disadvantage of conventional CCD image sensors designed for interlaced scan video systems lies in their low vertical resolution (see Scanning). To reach both high vertical and dynamic resolution, one needs progressive scan CCD camera. A way to grab both odd and even fields at the same time is to use a dual channel camera (for example SONY XC-7500/8500CE or compatible).
Such a camera is equipped with two video outputs, each of them providing conventional interlaced image (this is allowed thanks to newly developed CCD design - the chip contains two horizontal shift registers, one outputting odd, the other even lines to accordant video output).
While one of the registers provide just odd and the other just even field data every 1/60 or 1/50 sec, the outputting image signal of both video outputs is compatible to common EIA or CCIR standards. However using appropriate framegrabber, they can be formed to one full frame (i.e. 625 or 525 lines) high resolution image (see image bellow). Thanks to the fact, that both fields were read at the same time (during a single exposure), no motion blur will be caused.
Following pictures of Sherlock Holmes may help you understand, how two outputs of a dual-channel camera can be mixed to acquire full frame image.
| Original image | ||
|
||
| Video out 1 (odd field) |
Video out 2 (even field) |
|
|
|
|
| Video out 1 + 2 (mixed in framegrabber) |
||
|
||
Employing an electronic shutter, the camera CCD integration time (exposure) can be controlled to less than 1/60 or 1/50 sec to reduce smear when capturing fast moving objects. All light sensitive sensors can be simultaneously erased (while darkened shift registers are not affected). Effective time of exposure (the time between erasing the sensor and start of the read out process) can thus be freely controlled by the camera electronics.
Note: Unless using a progressive scan camera, one shutter grabs just one half an image (one field). It means that in case of composing full frame image, both odd and even fields will be shifted each other (the odd field of real image is captured at a different point of time than the even field is).
One of the two equal but vertically separated parts into which a common video (interlaced) frame is divided in an interlaced system of scanning. The odd field consists of 1-3-5... lines, the even one of 2-4-6... lines.
The point at which rays of light converge for any given point on the object in the image. Also called the focal point.
The total area, occupied by the television picture, which is scanned while the picture signal is not blanked. In conventional interlaced systems, the frame consists of two fields.
Computer card that samples and digitizes analog video signals so that the information may be processed, stored, or operated on by the computer. It is also called image acquisition or image capture board.
An increase in voltage or power, usually expressed in decibels.
A numerical value, or the degree of contrast in a video image. In the chart describing relation of real response (optical density) d of the CCD (or other medium) versus the logarithm of incident light intensity w, gamma represents the slope of linear part of that curve.
|
|
|
| Charge describing the response of a CCD sensor |
Ideal response of a sensor (gamma = 1) < /FONT> |
Abbreviation of Hole Accumulated Diode. A semiconductor structure developed for Sony third-generation CCDs. It permits a considerable increase in pixel count and improves overload and vertical smear characteristics.
IT sensor without lenses:
IT sensor with on-chip lenses:
An aperture of adjustable size, normally forming part of the camera lens, which is used to control the amount of light reaching the CCD. Iris control may be either manual or automatic, depending on the application / type of camera.
Small, rapid variations in a waveform due to mechanical disturbances or to changes in the characteristic of components. They are caused by variations in supply voltages, imperfect synchronizing signals, circuits,etc.
A transparent optical component consisting of one or more pieces of optical glass with surfaces so curved (usually spherical), that they serve to converge or diverge the transmitted rays of an object, thus forming a real or virtual image of that object. Often used in groups for light control and focusing.
There are several standards for mounting lenses to a camera.
Typically a table with 256×2 fields, intended for adjusting 8-bit grayscale or paletted color images. The color of every pixel is found in the first column of the table and replaced with accordant color from the second column. Binarization, gamma correction and many other operations may be easily performed by means of a LUT.
Following image illustrates usage of a LUT for grayscale image binarization with a threshold of 127.
Minimum amount of light (in luxes) that is needed to generate reasonable image signal in accordant CCD chip.
Irrelevant or meaningless data resulting from various causes unrelated to the source; random, undesired video signals. The word «noise» originated in audio practice and refers to random spurts of electrical energy or interference. In some cases, it will produce a «salt-and-pepper» pattern over the video picture. Heavy noise is sometimes referred to as «snow».
The noise of an image can be reduced for example by averaging several consequent pictures (of a static image) or by use of smoothing filters.
In the partial scanning mode, the camera reads out only a limited number of lines, not the entire CCD sensor. The vertical resolution is thus decreased, but the output frame rate is accordingly increased. The partial scanning mode is used in applications requiring a high speed camera output, when resolution is not a critical parameter.
The following image shows an example of partial scanning chip. Besides the effective lines (containing a real video information), one has to count with the vertical blanking period same for all scanning modes.
PCI local bus is a standard used in today's computers for high speed component-to-component connection. It offers features like bus mastering, DMA, data bursting, scaleability, plug&play support. Thanks to its high performance (up to 132 MB/s throughput), the modern framegrabbers can offer real-time transfer of video data to the main and video memory.
Short for Picture Element. A pixel is the smallest area of a television picture capable of being delineated by an electrical signal passed through the system of part thereof. The number of picture elements (pixels) in a complete picture, and their geometric characteristics of vertical height and horizontal width, provide information on the total amount of detail which the raster can display and on the sharpness of the detail, respectively.
Pixel binning technology used by some CCD cameras lies in combining adjacent pixels with the goal of faster output when needed. For example when pixels are vertically combined to pairs, two times higher frame rate may be achived (with 1/2 vertical resolution).
New CCD design, allowing acquisition of both even and odd fields at the same time. In this type of sensor there is no integration of signals between adjacent lines, so each pixel contains information from one complete frame.
While progressive scan method provides signals from all pixels during a single exposure, it is possible to obtain image signals delivering both high vertical and dynamic resolution without a mechanical shutter, which is impossible with a conventional CCD. The progressive scan CCD is ideal for use where rapidly moving objects must be captured with high resolution, such as in image measurement and image processing applications.
In the following table you can compare the results of shooting moving objects by a progressive scan camera with results of a conventional CCD camera. The results are different for vertical and horizontal contours and for various types of move. We used image 100×100 pixels.
| Original image (full frame) |
Odd field | Even field | ||
|
|
|
| Objects' move between capturing two consequent fields | Conventional CCD | Progressive scan CCD | |
|---|---|---|---|
| Field integration mode | Frame integration mode | ||
| One pixel horizontal | 
|
|
|
| One pixel vertical | 
|
|
|
| No move | 
|
|
|
| One pixel both directions | 
|
|
|
The term «resolution» appears in two slightly different meanings:
The resolution in the second sense is a function of the first meaning of the word and of the pixel depth (number of colors used in current color model).
Note: Besides the spatial meaning of the word, term «resolution» is also used in conjuction with the number of colors in the image. Color resoulution is just another name for the pixel depth.
The Restart/Reset function is a feature of some CCD cameras, that enables to start new image integration at any time (specified by external reset pulse) and only those pictures at the specified timing to be recorded in the picture memory. The major difference from a trigger shutter mode is that Restart/Reset function is not a shutter mode, the integration time is fixed: 1/60 or 1/50 sec, depending on the video norm.
The common mode of the Restart/Reset function causes a new integration start after supplying the reset pulse (the pulse resets internal vertical sync timing of the camera). While one field reading time (1/60 or 1/50 sec) is needed for the charge integration, irrelevant signal is output during this interval. Shooting information accumulated within the first field after the reset is output in the second field and so on (in frame integration interlace mode 1/30 or 1/25 sec the Restart/Reset requires two vertical sync pulses for integration, meaningful signal is output since the third field).
An other kind of the Restart/Reset function (which requires further internal setting in the camera) assures the slow speed shutter operation. In this mode, the first external trigger pulse resets the camera vertical sync timing. The camera is kept on standby for a prolonged period of time after this, integrating new image. Nothing but noise is output from the camera now. The second trigger pulse ends image integration; all the signals integrated since the previous reset are taken out as standard video within next two fields. This operation may be used for capturing dark objects.
The process of read out the contents of a CCD chip. In opposite to new progressive scan cameras, common CCD sensors can read just one half an image in one step (just one half of lines present in the sensor can be output at the same time). It can be performed either in interlace or noninterlace mode.
Note: both images illustrate the case of EIA video norm.
The sensitivity of a camera is defined as the lens iris aperture required to provide a video signal of standard level at a specified lighting level. In general, sensitivity is measured using 89.9% reflectance grey scale chart illuminated with 3200 K illuminant at a level of 2000 lux. Often also specified as lux level for fixed lens iris aperture.
The ratio of the value of the signal to that of the noise (expressed in decibels). The higher the number, the better.
Special mode of picture integration similar to the «B»setting on the shutter of a still camera. Effectively allows a CCD camera to reach a very high sensitivity level on non-moving objects. It requires external control of the camera and a frame memory. Also known as long-term integration.
A term for a complete vision system contained in the camera body itself, including imaing, image processing and decision making functions. While the common smart cameras are intended just for the dedicated systems, the latest PC technology enables development of devices fully compatible with desktop PCs. This category of smart cameras thus provides a standard API and thus much wider functionality.
Pixels of the same x and y dimensions (pixel aperture ratio PAR = 1). In case of rectangular (non-square) pixels (usual in TV) one have to maintain the aspect ratio when measuring objects, because the dimensions of stored frame aren't equal to true dimensions; resolutions along x and y axis aren't the same. Use of square pixels solves such problems - picture elements are equally arrayed in both directions, and allow easy addressing. Thus aspect ratio of the image does not require adjustment. This is needed in image processing tasks requiring accurate image measuring.
An image captured by chip with rectangular (non-square) pixels; pixel aperture ratio is not equal to 1 (in this case PAR > 1):
The same picture in memory sampled without pixel aperture correction (left) × with pixel aperture correction, oversampled (right):
Displayed memory; resulting aspect ratio deviation ARD <
1 (left) × ARD = 1 (right).
The structure of a basic CCD has the disadvantage that the charge on individual sensors is corrupted by changes in illumination during the read out process. Today's CCD sensors are developed with two different structures with two different ways of read out process and with slightly different features.
|
|
|
|
|
|
Every CCD camera should be equipped with an internal circuit generating its internal timing. This information is used for synchronizing the read-out process from CCD chip and (in case of composite video signal) than provided to output. Except the internal sync, some cameras can accept sync information from an external device (for example from a framegrabber) as well. Such external sync signal is used to control a PLL/VCO, which provides the internal clock cycle. There are several types of sync signal:
Trigger is a special signal provided to a camera (or other device) to ignite some needed operation - most frequently to reset its timing, turn on a flash or a strobe etc.
The term «trigger» is sometimes used also in meaning of a trigger shutter (a shutter mode with random timing or even with random shutter speed). Such a randomness is controlled by the trigger signal mentioned above.
Transistor Transistor Logic, a digital signal using 0 V to 0.4 V to represent logical «0» and 3 V to 5 V to represent logical «1».
Below you can find descritption of some of the units frequently used in the camera/framegrabber technology.
Vertical smear is a phenomenon peculiar to some types of CCD camera which occurs when a bright object or light source is shot with the camera. This phenomenon is observed on the monitor as a vertical streak above and below the object or light source.
Reading out from a CCD sensor is regulated by a video norm that defines the timing and the level of the transmission. There are two different norms (both strongly related to an interlacing technique).
| CCIR | EIA (RS170) | |
|---|---|---|
| vertical reading frequency | 50 Hz | 59.9 Hz |
| vertical reading time | 20 ms | 16.68 ms |
| vertical return time | 1.536 ms | 1.271 ms |
| horizontal reading frequency | 15.625 kHz | 15.734 kHz |
| horizontal reading time | 64 µs | 63.55 µs |
| horizontal return time | 11.52 µs | 10.76 µs |
| number of lines | 625 | 525 |
| number of visible lines | 574 | 485 |
The charge accumulated in the CCD sensor is used to form resulting video signal by camera electronics. Resulting waveform should correspond to according video norm.
Color video signals ae composed of luminance and chroma (color) information. Composite signals carry both parts on a single line (wire), whereas component signals (Y/C, RGB) carry the video information separately on more lines. Breaking up the signal components generally improves signal fidelity, especially when recordig or balancing color.
A function enabling adjustment of the image colors to make the white objects really appear as white. Thus one can avoid color shifts caused e.g. by different illuminating conditions.
A compound lens which remains in focus as the image size is varied continuously. May be motorized or manually operated.