AGOCG logo
Graphics Multimedia VR Visualisation Contents
Training Reports Workshops Briefings Index
This report is also available as an Acrobat file.
  Back Next Title

Section 2: Image Capture

Image generation can be handled in a number of ways. Spreadsheets can be used to store data and generate graphs and charts. Line diagrams, flow charts and sophisticated graphics such as molecular models and chemical formulae can be generated using graphic design and drawing software programs. Photographs, transparencies and documents can be scanned or captured using a camera and image capture board. Alternatively, computer-ready images, for example, samples of clip art, may be imported into computer-based learning material.

This following section will focus on the capture of photorealistic images (transparencies, prints, objects, etc) for computer-based learning materials.

There are several ways of capturing images but all have one thing in common; the visual information needs to be converted into an electronic signal before it can be stored, edited and ultimately displayed. We perceive images by the light that reflects off or passes through an object. This light or optical signal is captured and converted to an electronic signal which can be stored. The electronic information or data can be stored as either an analogue or digital signal.

  • Analogue: For analogue data the signals are based on the variations that occur in a continuous electronic signal. Such signals occur on cassette tape, videotape and videodisc. Such signals can contain an infinite range of shades, tones and colours and as such transparencies and prints can be referred to as analogue or continuous images.
  • Digital: With digital data the electronic signal is represented as a '1' or '0', that is, on or off, logical yes or no. Digital data are stored as bits and bytes and examples are the data stored on a hard disc, floppy disc, compact disc or DAT (digital audio tape).
The process of image capture can be divided into four steps. These are:
  • Capture
  • Store
  • Edit
  • Display

Capture and store will be considered together as will edit and display. For each group the advantages and disadvantages of the analogue and digital scenarios will be discussed. Hardware and software issues will also be explored.

Image capture and storage:

The analogue model

This sort of model is not for the individual wishing to capture half a dozen or so pictures. However, it is an essential consideration when archiving material such as thousands of transparencies and prints, for example, for archiving museum specimens, works of art, medical images, etc. The hardware is expensive but in many instances, such as those referred to below, can offer a cost-effective medium.

This model has been used to produce a number of videodiscs such as the Bristol Biomedical Videodisc (Williams, 1992) and the videodiscs produced for the Wellcome Trust Tropical Diseases Videodisc Project (Longstaffe et al., 1992).

Figure 2.1 shows the recording studio for videodisc production at Bristol University.

Figure 2.1 The Laser Videodisc Recording System at Bristol University.

The recording line comprises: a carousel unit (1) inside which is mounted a 3-chip broadcast quality camera; the camera control unit (2); two colour monitors (3), a black and white camera (4), a wave form monitor (5); a colour balancer (6); the Sony Laser Videodisc Recorder (7) and a computer running the cataloguing software (8). The wave form monitor is used to monitor the quality of the signal and colour correction is carried out using the colour balancer. The corrected image is viewed on the second colour monitor prior to recording. The recording process is under computer control; the frame numbers are automatically inserted into the appropriate database records.

The equipment or hardware required includes:
  • camera (input device)
  • colour balancer (restore colour balance to images)
  • wave form monitor (monitors the quality of the signal)
  • colour monitors (to visualise the process)
  • video tape recorder/recordable videodisc (recording device)

Not all of the above equipment need necessarily be bought. It will depend very much on what it is you want to do. It is best to take advice from a reputable dealer. However, you will need a camera and a recording device. The camera may also double up for digital image capture (see the digital model).

In addition to the above list of hardware you will also need some device to support the image that is being captured. This may simply involve projecting the image (e.g. transparency) onto a white screen and pointing a video camera at the projected image. The quality of this captured image can be improved by projecting the image directly into the camera. Figure 2.1 shows a piece of equipment known as a Kindermann Unit. This houses a normal carousel unit but the camera head has been mounted inside and the optical signal directed into the camera lens. Alternatively, the camera can be mounted on a stand (rostrum) and a light source placed above (in the case of objects and prints) or below, e.g. a light box, (in the case of transparencies). In this instance a number of different lenses are required. You may find that a lot of this equipment already exists within your institution.

It cannot be stressed strongly enough that the quality of the captured image will depend on the quality of the original. The optics and electronics of the camera will also affect quality. Video cameras contain photoreceptors known as charged couple devices (CCDs or "chips"). It is these CCDs that convert optical information into electronic data. The number, resolution and spatial luminance of these CCDs will affect the quality of the resulting image. Video cameras are typically single-chip or three-chip. Three-chip cameras are generally thought to be of higher quality, the chips separately capturing the red, green and blue parts of the light signal. However, there is much debate and many arguments for and against the purchase of three chip cameras over single chip cameras. Certainly more technical support in needed with three chip cameras; as the chips need to be kept in alignment, any spatial misalignment will show on the captured image. Good-quality images can be generated from high-resolution single-chip cameras. Whichever choice of camera you decide upon make sure the camera has the capability of outputting an RGB signal (see section 3, results and conclusions).

Whatever camera you finally choose, it is important to remember that the better the image you start with, the better the resulting captured image.

The resulting tape or recordable videodisc containing the images can be used as a 'one-off' or alternatively be sent to a videodisc manufacturer and be mastered and replicated to produce the required number of standard read-only videodiscs.

The digital model:

To convert an analogue image (whether transparency or stored on tape/videodisc) to a digital image, a sampling process occurs that creates discrete finite picture elements and assigns to them both a location and a colour or grey value. This process is known as digital imaging or digitising.

Video capture is the process of capturing and displaying sequential images rapidly in real-time. Single frames can also be captured from video if this is your source material

There are three possibilities as far as equipment or hardware is concerned:
  • Digitising board (frame grabber/video board) plus analogue input (e.g. camera, videotape)
  • Scanner (hand-held, flat-bed, transparency)
  • Digital camera

Digitising boards: are boards or cards that fit into the expansion slots of the computer, allowing the capture and storage of a digital file. Image capture devices that capture images via the parallel port of the computer are also available. All need an analogue input either directly from a camera or from a videotape, videodisc or other analogue storage medium. Again, the better the quality of input, the better the resulting digitised image. A variety of digitising boards are available for both PC and Apple Macintosh platforms and range from as little as 79 to 4,000. The price reflects the number of options offered and its capabilities of allowing RGB, S-video and composite inputs. The top range boards will also provide video output, allowing the recording of computer presentations and animations to tape or other analogue devices.

Some image capture boards come complete with software that allows not only image capture but also extensive image manipulation. Other boards allow capture only. Extra software is then needed to take that image and alter it further. This is important to remember if your budget is limited.

Some capture boards, known as video capture boards, allow the capture of full-motion video using a video camera or VCR . This involves a great deal of storage (2-10 MB for every 1 minute of video) and compression methods need to be employed (see sub-sections on storage and compression). The cost of video capture boards generally reflects the frame rate and size at which the video is captured.

For more details on image capture cards see Section 3.

Scanners: These range from the low-end hand-held scanners to the high-end drum scanners. Between these two extremes lie flat-bed and transparency scanners that are more commonly used for digital image capture and for Desk Top Publishing applications. Some scanners on the market today combine flatbed scanners with a transparency module, that is, they accept transparency and reflection copy (Rogers, et al., 1993). Alternatively, dedicated slide scanners are available. Scanners may be monochromatic or colour . The size of image that can be scanned varies from scanner to scanner. Some accommodate originals, from 35mm up to A4. Scanners specialising in capture of X-rays are also available. Many of the more popular computer magazines regularly include surveys and tests of the latest scanners.

However, you do not need to purchase a full-size flatbed scanner if you are not intending to use large images. Small inexpensive hand-held scanners (both monochrome and colour) are adequate for images such as clip art, small diagrams and photographs.

Again, charged couple devices are used to create an electronic signal (voltage) that is proportional to the amount of light it receives. An analogue to digital converter converts the electronic signal from the photoreceptor to digital form. The size of the CCD, its sensitivity (number of cells), the sampling rate and repeatability of the scanning motor are all important in determining the resolution and quality of the scanning process. The CCD consists of a horizontal array of cells or elements. There are CCDs with 2,400, 3,200 or 4,800 elements. It is these that determine the optical resolution (dpi - dots per inch) in the horizontal. The optical resolution in the vertical is determined by the stepper-motor. Optical resolution can be artificially increased by mathematical tricks or interpolation. That is, the resolution is increased by adding artificial dots between the dots gathered by the scanner. Interpolated data is only as good as the original information and only makes the image worse when the image is of poor quality to start with. High resolutions are only needed for those images that need enlarging, for example, transparencies, or to meet the needs of professional graphic artists. If this is needed be aware of the true dpi or resolution of the scanner before purchasing.

If purchasing a scanner for capturing images that will eventually be outputted to print then the resolution of the image should be determined by the output device. Scanning at higher resolutions will only waste valuable time and disc space. Redundant data will be disregarded by the printer and there will be no benefit in image quality. If outputting to print the majority of applications will be handled easily with 300 x 600 dpi optical resolution scanners.

The new scanners appearing on the market today are offering 36 bit colour, that is 12 bits each for the RGB components. However, very few applications currently support 36bit images.

Once the original has been mounted ready for scanning the operator can prescan the image and, depending on the software employed, can rotate, crop and resize the image, mark the highlight and shadow areas, select a specific area and sometimes perform colour correction on the image. Some scanners have plug-in software modules that drive the scanner directly from some of the more popular image processing software programs on the market. These packages provide excellent manipulation facilities and many special filters and effects. However, complex image manipulation may require an investment of time for learning.

Digital cameras: Like conventional 35mm cameras, digital cameras use optical lenses. However, rather than capturing the image to film, the image is captured with a CCD chip and stored internally either as analogue or digital signals. The images can be transmitted over networks and telephone lines, displayed on the screen and printed. The speed with which images can be produced and transmitted is of great value to people who need images fast. In instances where the images are stored as analogue, one still has to plug the camera and disc into a digitising board. The images once captured can then be altered in the appropriate way.

The first generation of digital cameras, such as the well known Xap Shot or Ion camera from Canon, do not have the resolution to compete with conventional photographs. Images are stored as analogue signals on small discs, each disc taking 50 images. Vast improvements have been made in this area; many of the major manufacturers, such as Canon, Kodak, Minolta and Nikon, are producing high resolution digital cameras based on SLR cameras. The newer digital cameras come equipped with internal hard discs or have utilised PCMCIA technology for storing images for later viewing and manipulation with a computer. The cameras can also be directly built into an imaging system and come with software plug-ins for the more common image manipulation packages such as Adobe Photoshop for capturing images directly. Digital cameras are expensive and the digital file is very large, up to 18 MB in some cases. The advantage of these cameras is their portability and the elimination of the need for chemical processing. For further information see Brown (1994a, b) and Robinson (1993).

PhotoCD: PhotoCD is a digital image format and provides a mechanism for transferring photographic film or transparencies onto compact disc at very high resolutions. PhotoCD images can be manipulated further using image processing software. For further details see section on PhotoCD.

In summary, choose a digitising method that takes into consideration the source material (videotape, transparency, etc) and target display. See also section 3, Conclusions: Recommendations and Guidelines for Image Capture

Before digitising any material be sure to have read or be aware of any copyright restrictions and if necessary seek permission first.

TWAIN TWAIN is a protocol which defines how bitmapped images from scanners and other image capture devices can be captured directly into image processing packages. For example, image manipulation software such as Aldus Photostyler, Adobe Photoshop and Micrografx PhotoMagic are TWAIN aware and can directly capture images from source. TWAIN drivers are supplied with the image capture hardware.

Cataloguing/indexing images

When capturing images and storing them on analogue or digital storage devices, one area that is often forgotten is that of cataloguing. With analogue systems, the images are often stored and referenced as time-codes or frame numbers. With digital systems one is limited to a filename of a few characters (eight letters on current IBM compatible PC's) which is often insufficient for immediate recognition. Imagine you have several thousand images stored on a compact disc and you need access to those that illustrate a particular object or scene. Without an index, your only choice is to load each image sequentially. You may only require a few images for a tutorial but very quickly the images become unmanageable. At the very least one should keep a written record of the frame reference/file name with corresponding key words. However, once into the hundreds, it becomes essential to keep a proper database/index of the images. Information on the images that one might wish to keep, especially if coordinating an image archive involving several donors, will include sections on the details of the donors, copyright information, image subject matter (description, key words), frame number/file name, physical location (e.g. catalogue videodisc no. 2/CD-ROM no.7, subdirectory Landscapes) and anything else particularly relevant to your collection.

Again, there is much software available to help. Database tools provide a means of indexing data (still images, video, animations, graphics, audio, etc) in a form that facilitates easy retrieval. They incorporate tools that allow the user to create, update, interrogate, relate and build temporary image sets from the main database.

One might choose to use a relational database or cataloguing software that specialises in search and retrieval. The advantage of the latter is that searches of the database can be carried out on any word or piece of data. However, the drawback is that the image you require may be there but the data or word on which the search was performed is not present. To help overcome this many people catalogue using a classification system. In addition guidelines for writing details of the description and key-words should be provided if your images are coming from a number of people. Different people will describe the same image differently. A well designed database will accommodate all these views. Depending on the size of your project, for example if creating a resource of images, and your access to programming expertise, several search and retrieval mechanisms can be employed to assist in searching such as providing a list of headings (this may be generated from any classification system used), or key words available, spell checking, synonym lists, context based searching, etc.

Within the Educational Technology Service of Bristol University, two different databases are used for cataloguing image based resources; an administrative database based on the relational database model and a delivery database using search and retrieval mechanisms. By using the relational model for administering images, details such as the donor and institution, for example, need only be stored once for a large number of images. Only information needed by the users is then delivered with the images via the search and retrieval programme (Williams and Hammond, 1994).

In the case of digital images and other digital files, modern database software allows a record to be linked with the corresponding image file. Some software will display a series of the indexed images the size of a postage stamp. A number of commercial specialist image database software packages are now available on the market which will display your images as 'thumbnail sketches' and provide space for key words. Kodak and Aldus have developed such archive management software, 'ShoeBox' and 'Fetch' respectively. Both have facilities for tagging each image with numerous keywords and descriptions and are extremely easy to use.

For videodisc resources, some database software will communicate with the videodisc player provided the appropriate drivers are available to display the image on a second screen or, if using an overlay board, on the same screen as the computer data.

Back Next Title