The Software Revolution
Eight years ago the Asian Institute of Technology, at the forefront of remote sensing education in the Asia Pacific region, had a variety of image processing software, ranging from mainframe to PC.
Table 4 1986 Image Processing at AIT
| |
PC software |
Mainframe software |
| Price/seat |
$20000 |
?? |
| Special equipment |
$3000 |
$5000 |
| Image size |
512*512 |
512*512 |
| Display depth |
24 bits |
24 bits |
| Speed |
Slow |
Unpredictable due to competing
users |
| Cost of operation |
0 |
At least $1000/month |
| Ease of use |
Primitive menus or verbatim commands
No on-line help
Any user error required
restart
Cryptic, idiosyncratic terminology
Difficult to learn
Poor documentation |
With these software resources (which were generous for that period), we typically spent two thirds of the time allotted to digital analysis laboratory simply teaching the mechanics of using the software. Time available for students to actually do image analysis was correspondingly reduced.
Our own DRAGON software was developed partly in response to these conditions. By contrast, the same criteria applied to the current version of DRAGON
are shown in Table 5.
Table 5 1994 Image Processing with DRAGON 4.1
| Price / seat |
$400 to 995 |
| Special equipment |
None required |
| Image size |
1024*1024 |
| Display depth |
24 bits |
| Speed |
Fast |
| Cost of operation |
0 |
| Ease of use |
Full mouse-and-menu, or command-line or script-file-driven
Full context-sensitive on-line Help
(Help and menus in English, Spanish, Indonesian/Malaysian)
Tolerant of user errors
Industry-standard terminology
Easy to learn
Extensive documentation, carefully written for clarity and consistency
|
Goldin Rudahl systems pioneered the use of standard personal computers for remote sensing analysis, especially for education with DRAGON/ips. The goal was to provide low cost, easy – to – use software on a low cost, readily available hardware platform. Now, seven years later, a number of other companies have followed our lead. The significant point here is that the state of the art has changed dramatically. The level of performance outline above is what a remote sensing educator can and should reasonably expect of a modern remote sensing software package.
Internationalization – The Next Step in communication
Those who are reading this paper presumably understand English, and it may be supposed that all of your students have to learn one of the major technical languages. Nevertheless, a person is a most always most comfortable using his or her native language, and software which can interact with a student in familiar terms will be easiest to learn.
New software methodologies are now coming into use which facilitate the development of internationalized software –software which interacts with the user in a national language of choice. The problems of translation, of finding suitable technical terminology and suitably knowledgeable translators, remain the same. However, correctly designed software can no integrate translated prompts and message with a minimum of effort and without requiring any knowledge of the language by the software developers themselves. Our own newly available Bahasa Malasiya/Bahasa Indonesia version of DRAGON is an example of this process. (we also have a Spanish version; other similar projects are under way)
Perhaps even more important than the comfort of the student, however, is the need to communicate analysis results to the non expert : to government officials, funding agencies, and the lay public. Colorful and attractive maps and images are an important part of this, but clear labeling in a language understood by the intended use is crucial to communication. Especially here in Asia, this labeling involves not only the correct words but also use of the correct character set (or sets).
The apparently simple task of typing the name of a city, to appear as a label on the image, can take on considerable complexity. While easy enough in English, Indonesian, or Swahili, it becomes a bit more difficult with the accented characters in French and Spanish, quite a bit more difficult with alphabetic languages not based on Roman characters, such as Cyrillic, Greek, and Thai, and a major challenge when the character sets number in the thousands of characters. Of course, all these cases must be handled without compromising the simplicity of standard Raman character entry.
Recently, a new standard for representation (almost) all the world’s languages has been completed by the International Standards Organization working with the Uniceode Consortium and national groups worldwide (but especially in China, Taiwan, Japan, and Korea). New software, including DRAGON, is beginning to incorporate this ISO-10646/Unicode standard as the basis for a powerful and flexible text display capability (Figure 3).
Graphical User Interfaces
Most readers are probably familiar with graphical use interfaces (GUIs) such as Microsoft Windows, the Macintosh interace OS/2 Presentation Manager, and X Windows/Motif. AGUI extends the vocabulary available to the software designer for communicating with the user. Simple text messages are augmented by lines showing grouping or flow, small pictures depicting recognizable actions, standardized or conventional symbols, and dynamic displays which respond to user actions.
Graphical user interfaces are often hailed as being much easier to learn and use than “old –fashioned” character-based interfaces. When used well by the software designer, a GUI can facilitate the user’s comprehension of “what the software is doing”. When designed inappropriately, a GUI cumulated with proper use of the software, encourage a non-professional video-game approach, or even hide the inadequacy of the software.
One negative aspect of GUI is that it almost always requires more powerful hardware to achieve a particular level of performance than does a comparable non-graphical package. While this fact may become less important as hardware costs drop, it is still a consideration in examining options for educational software.
Acquiring and Managing Image Data
CD-ROM
An outgrowth of the audio CD, the CD-ROM may ultimately have as large an impact on history as the invention of printed books. Although less convenient than a book, CD-ROM is much smaller (10 cm3 vs 1200 cm3), contains much more data (6x 108 characters vs 6 x 106), costs much less to manufacture ($1 vs$5 , in large quantities), and has about the same durability.
Faced with this dramatic alteration in the economics of data, government agencies and corporations are realizing that it is cheaper and easier to publish their results as a handful of CD-ROM disks than as half a ton of paper-and then to relieve budge pressures by pricing the disks low and selling as many as possible. Thus it is now possible to buy, for US$100 or so per disk, detailed geophysical, climatologically, ecological, demographic, and historical data sets covering an entire country.
Vendors of earth satellite data, too, are beginning to use this technology. The data sets themselves are still costly, but CD-ROM data from SPOT Image and EOSAT, for example, can be read on any PC equipped with in inexpensive CD-ROM drive.
These developments are particularly significant for remote sensing educatin because they provide direct access to large data sets. Unitl the advent of CD-ROM data distribution, there was no practical or affordable way for educational labs equipped with personal computers to read and process data in the form distributed by vendors. Tape drives were too expensive and complex for most educational environments. Educational users of remote scenting data typically had to find some organization with a mini-commputer and a tape drive, to read their data and subset it into pieces small enough to be transferred to a PC hard disk. This was slow, cumbersome, and error-prone.
CD-ROM technology makes it possible for a personal computer user to get fast, immediate access to enormous amounts of image data. Furthermore, CD-ROM, unlike tape, is a random access medium This means that a user can select exactly which bands or sub areas of an image he or she wants to process, and quickly start work on this data. With traditional sequential access media like tape, it might be necessary to read past dozens of megabytes of unwanted data, in order to order to reach and extract a subarea of interest.
One of a kind CD-ROMs
The newest twist on CD-ROM technology is the ability to create your own. Special equipment is required to produce a CD-ROM, but service bureaus are beginning to appear which do just that, at prices starting as low as US$125 for the first copy and dropping for several or a dozen copies. The advantage, of course, is that these CD’s can be read on any standard CD-ROM drive.
It is too early to tell what impact this exciting development will have on remote sensing education, but one can imagine a variety of applications. For instance, regional organizations might publish an atlas of local imagery for use in schools and universities. The results of student projects might be captured on CD-ROM and disseminated to other educational organizations.
Image Data from Hardcopy or the Real World
Remote sensing is sometimes assumed to mean processing of data from earth-orbiting satellites. However, there are times, when satellite data is too expensive, or too low resolution, or simply unavailable.
Aerial photos constitute an inexpensive source of image data which can be digitized and subjected to digital analysis. The spatial resolution of aerial photos is much higher than satellite data and the photos may be available inexpensively from government sources. Alternatively, it may be possible to commission special flights at reasonable cost. Furthermore, aerial photo archives frequently stretch back much further into history than satellite image archives. This makes aerial photography particularly useful for multitemporal work. Color-Infrared photos even provide some multi-spectral information.
One problem in working with aerial photography is the fact the high spatial resolutioin can overwhelm processing capabilities when working with a large area. Photographs can be successfully scanned at quite high resolution, but the resulting images scan be much larger than can be viewed or conveniently processed.
Other technologies exist for capturing images of the real world directly, without going through a hardcopy photography phase.
- A digital camera functions much like any photographic camera, but the image produced is in digital form and can be downloaded directly into a computer.
- A frame grabber captures a video image, either form video tape or form a video camera, into the computer in image form.
At present, both these devices are too expensive and provide insufficient spatial resolution to be generally useful in remote sensing education; however, when live images must be captured, these may be the most appropriate way. In future, we can expect them to become more useful.
Remote Sensing Education – Communicating the Need
Remote sensing is not an end in itself. It is a tool a very powerful one for understanding and hopefully improving our natural and social environments. To those of us here at ACRS, this is self-evident. To the government ministers and officers who control our budgets and implement politics, helpful or otherwise, it is not at all self-evident. Nor should it be: their function is to require that we demonstrate the value of remote sensing and remote sensing education.
Remote sensing technology is not very accessible to the lay person. Necessar foundation concepts such as electromagnetic radiation, multispectal analysis, orbital paths, and so on, are difficult and unfamiliar. Further more, it is not easy to relate the highly technical field of remote sensing to real-world problems like hunger, disease, and urban crowding. Yet, we need to make these connections in the minds of decision-makers, if we are to ensure ongoing support for education in remote sensing and associated technologies.
Recently, the most ubiquitous technology of our era, television, has provided some assistance in communicating the potential importance of remote sensing. Educational and news programs have started to use remotely-sensed images to help viewers visualize phenomena at a regional scale. For example, a widely-publicized image of the Rodonia region in Brazil shows more dramatically than any characteristics the extensive damage being inflicted on the Amazon rainforests.
In most cases, television viewers are probably not aware of the source of these images. However, this type of usage makes the general public and the decision-making community more familiar with the products and applications of remote sensing can do, and why it is important.
The expanding use of Geographic Information Systems can also help to justify support for remote sensing. Increasingly, appropriately or not, remote sensing is being seen as a component or sub area of GIS. Possibly it is easier to explain geographic analysis to a layperson in GIS terms, by focusing on the analogies to conventional maps. Sine most people are at least somewhat familiar with thematic maps, it should be fairly straightforward to illustrate how GIS can be used to understand environmental and social problems.
Thus, to increase support for remote sensing education, we should be explicitly broadening our educational objectives to include other aspects of spatial will make it easier to communicate the need for educational programes to governments, funding agencies and the general public. At the same time, revising our definitions in this manner will allow us to provide students and trainees with remote sensing education broader, deeper, and much more relevant to the world’s problems, than what we have offered in the past.
We, the technical experts, need to recognize the importance of educating the general public about these technologies for the sake of society as well as to ensure support for formal educational programs. Geographic knowledge and environmental awareness are essential for citizens making decisions n today’s global society. Although discussions of remote sensing education have typically focused on programs at the university level and above, we are beginning to see efforts at the secondary and even the primary school level.
It is particularly critical at this time to focus on remote sensing education in the so-called less developed countries. Remote sensing is a discipline which was originally sponsored by a few large industrialized countries. Now, however, foreign aid is becoming unreliable due to worldwide financial inhabit, and developing countries re beginning to question the motives of externally sponsored aid projects. At the same time, developing countries have the most to lose from environmental crises and mismanagement of their national resources. Therefore, remote sensing must become a respected and economically viable activity within each individual country. Policy advice based on remotely sensed data must originate from and be supported by local expertise, locally educated.
Conclusions
The task of providing realistic, relevant education in remote sensing has become less difficult, due to the technological trends highlighted in this paper. Adequate funding and in fracture remain a problem, but given even a modest budget, it is now feasible to offer training programs that are more comprehensive and advanced than anything we might have imagined a decade ago.
Increased public comprehension and awareness of the methods and purposes of remote sensing is seen as a key to public support, while attention to maintaining connections to related disciplines may guarantee a continuing demand for the products of remote sensing.
