ABSTRACT: The Distance Diagnostics through Digital Imaging project enhances the ability of the University of Georgia Cooperative Extension Service to evaluate and propose solutions for agricultural problems, including plant diseases and pests, through the use of digital imaging and the World Wide Web. Imaging stations consisting of computers, digital cameras, microscopes and image-capture devices have been deployed in 94 county offices and in 3 diagnostic labs. To date, over 1,200 samples consisting of more than 3,600 digital images have been processed. This paper will focus on the outreach educational benefits of this technology to Georgia citizens. Aspects of the project to be covered include a description of the Georgia Extension Service and an overview of the project, initial planning, stories, lessons learned, implications to distance learning, and plans for the future.

The University of Georgia (UGA), founded in 1785, is the oldest institution of higher learning in Georgia and was the first institution in this country to be chartered as a state-supported university. The UGA College of Agricultural and Environmental Sciences (CAES), originally the State College of Agricultural and Mechanic Arts, was formed in 1872 and today employs nearly 2,000 faculty and staff. The CAES is one of thirteen colleges that make up The University of Georgia, a land- and sea-grant university. The mission of the CAES is to seek, verify and apply knowledge related to agriculture and the environment, and to disseminate this knowledge through student education and public outreach programs. The Georgia Cooperative Extension Service (CES), in partnership with the United States Department of Agriculture, is the agricultural education outreach component of the CAES. The Georgia CES was established as a result of the Smith-Lever Act of 1914 and serves 159 counties - making it one of the largest extension services in the country. CES programming includes agriculture, family and consumer sciences, and 4-H and youth.

County Extension faculty members are located throughout the state to serve citizens in every county. For farmers, homeowners, plant nurseries, etc. who are in need of agricultural advice or information, these agricultural experts serve as the primary point-of-contact and the pathway through which specialized information may be obtained from CAES researchers, teaching faculty, and extension specialists.

One of many responsibilities of county professionals is the design, development, and delivery of educational programs to agricultural producers. These programs teach producers methods for maximizing crop production and improving sustainability while maintaining environmental quality. Faculty members also promote sustainable agriculture by providing up-to-date crop production recommendations through research-based, non-biased diagnostic programs. Departments within the CAES have set up a number of formal clinics and other facilities and methods for such purposes as soil sample analysis and insect, nematode, and plant disease identification. County faculty are responsible for relaying these samples to the appropriate diagnostic site as well as for communicating educational information, including the analysis and accompanying recommendations, back to the client.

Distance Diagnostics Through Digital Imaging (DDDI) is a three-year project that is funded primarily by a gift from a private foundation and by state lottery matching funds, totaling a little more than $1 million. The CAES Office of Information Technology and the Plant Pathology Department developed this project jointly to address the need for more rapid access to visual information. Such information may be used to aid in plant disease diagnosis, insect identification, or any number of other Extension applications.

The principal investigators, Mr. Robert D. Hamilton, Coordinator for the Office of Information Technology, and Dr. Edward A. Brown, Plant Pathology Extension Coordinator, championed the cause for the design, development, and implementation of this technology for several years before major funding became available. Dr. Brown began investigating digital diagnostic techniques during the early 1990s as such techniques began to gain acceptance within the medical community. Mr. Hamilton began demonstrating the power of the technology to the organization in 1994 as the necessary equipment became affordable and, therefore, accessible and feasible.

As was noted earlier, this project arose as a response to a specific need: rapid access to visual information for agricultural evaluations. Each year, Georgia farmers submit more than 4,000 diseased or insect infested plant samples to the UGA Extension Plant Pathology Plant Disease Clinic. Historically, these samples have been either mailed or hand-delivered to the Clinic. These methods of delivery can be costly in terms of both time and/or money. The primary drawback, however, may not be quite as obvious. Approximately one-third of the samples submitted to the Clinic each year are found to be inadequate, inappropriate, or have deteriorated en route making diagnosis impossible.

To reduce sample deterioration and speed delivery and response, County Extension faculty have been provided with and trained to use microscopes with image-capture devices, digital still cameras, and the appropriate computer equipment. Faculty members use this equipment package, known as "agricultural diagnostic imaging stations," to photograph field symptoms, insects, foliage symptoms, and other identifying agents such as weeds and crop patterns. Compound microscopes and stereoscopes, fitted with cameras and image-capture devices to facilitate digital imaging, are used for close-up examination of insects, diseased plants, weeds, horticultural material, etc. This technology emphasizes proper pest and host plant sampling.

Once captured, images are uploaded via the World Wide Web (WWW). Paper forms have traditionally been mailed along with physical samples. These forms are mimicked on the DDDI web site. Such information as type of sample (commercial or homeowner), grower name and address, grower type, plant name, crop area, description of problem, chemicals applied, county of sample origin, etc. are required on these forms, along with image files, for submission of the sample. The textual data are written directly into a relational database and the image files are transferred from the client computer to the DDDI server and archived upon sample submission. The image files are automatically renamed and then associated with the accompanying textual information in the database. These sample records are also automatically tagged with time and date stamps.

Once the user has committed the sample to the DDDI system, an electronic mail (e-mail) message is automatically generated and sent to the appropriate diagnostician informing that person that there is a sample available for identification. The diagnostician is designated by the system based upon responses provided by the user in key fields of the submission form.

The diagnostician, having received e-mail notification, searches for and retrieves the sample based upon the sample number provided in the e-mail. That person then evaluates the sample and formulates a reply. In the original system, the diagnostician had the responsibility of contacting the person submitting the sample via phone, fax, e-mail, etc. to deliver the diagnosis. The newer versions of the DDDI system allow the diagnostician to record the evaluation and any accompanying recommendations into the database. Responses are automatically linked to the original submission data, and an e-mail containing the evaluation information is then generated by the system and sent to the individual who originally submitted the sample.

Each DDDI agricultural diagnostic imaging station consists of a computer, a printer, a hand-held digital camera, a compound microscope, a dissecting microscope, a video camera, a still-frame video capture device, and a collection of plant disease diagnostic compendia. The video camera is connected to the computer via the video capture device. This camera is mated with an adapter tube that can easily be moved from one microscope to the other. This allows the user to capture digital images from either source. The hand-held camera is flexible in that it is portable, can zoom up to 10x, and will still handle macro (close-up) shots as near as 1 cm. For this project, we chose the Sony Mavica for its macro capabilities as well as its ability to save images directly to a floppy disk, greatly simplifying the process of transferring images from the camera to the computer. The complete DDDI system gives the user the ability to capture images in a visual range from 60x microscopic up to 10x telescopic.

The DDDI system was developed on and runs under Microsoft Windows NT Server 4.0. and its WWW services are provided by Microsoft�s Internet Information Server 4.0. Originally, the system was designed around the Corel Paradox database engine. This product was chosen due to the availability of experienced development personnel and because of its wide usage within CAES. Newer versions of DDDI are being developed using Microsoft Access. This move has given programmers visual development tools that were not available with Paradox. Also, Access allows multiple users concurrent access to data, and its databases are scalable. It is anticipated that the Access database will be upsized to the Microsoft SQL Server engine. This will give DDDI an "enterprise" level system capable of handling data stores in the multiple terabyte range.

DDDI web-to-database connectivity is accomplished through the use of Allaire�s ColdFusion Server 4.0. This product allows programmers to use the ColdFusion Markup Language (CFML) that closely resembles the commonly used HTML. This language allows programmers to write code to communicate data exchanges to and from a database and will even allow the manipulation of data already stored inside of a database. HTML and CFML can be integrated seamlessly onto one web page essentially increasing the final potential functionality of such a page.

The DDDI imaging stations were distributed to all 94 sites over a period of two years. This process was well planned and carefully executed. The stations were distributed in groups of five or six at a time, and training took place at the time of distribution. Since the stations were to be placed strategically around the state, each county that would receive the equipment was clustered with one or more surrounding counties that would not. Faculty representatives from every county extension office in the state participated in the training. Those faculty members not receiving an imaging station were trained alongside the representative that would take home the equipment for their respective cluster.

The training for each group took place over two days. On the morning of the first day, faculty members were greeted with a mound of boxes. That first morning was dedicated to guiding the group through unpacking, arranging, and connecting all of the equipment. The afternoon was dedicated to familiarizing the group with the different ways by which digital images can be captured using the system.

The second day�s training was geared toward teaching participants the proper methods for preparing samples. Techniques for capturing appropriate diagnostic characteristics were also discussed and demonstrated. The second afternoon involved the actual submission of test samples to the system. The group was then guided through dismantling the equipment and packaging it for safe transport to the county.

Again, this process was painstakingly planned and executed. It was important that everyone involved receive the same level and quality of training. Even those faculty members who would not have the opportunity to take an imaging system home to their respective counties need to be intimately familiar with the system�s functionality. Diagnostic time can still be reduced when these individuals travel to an adjacent county to use a DDDI imaging station rather than submitting their samples using traditional methods.

As a result of additional funding from CAES, all county extension offices in the state received hand-held digital cameras. Images from these cameras can be submitted along with textual information through any Extension computer having WWW access, even if no imaging station is available in the county.

We believe that DDDI serves as a cognitive tool to promote the ongoing and continuous education of our county faculty. While these faculty members are recognized as being highly trained and educated, and are considered to be "the local agricultural expert", the importance of continuing professional growth is imperative to their success in supporting agricultural activities of the citizens of Georgia.

The body of knowledge pertaining to current agricultural markets, commodities, techniques, etc. is gigantic in size and continues to grow at a staggering rate. Georgia alone has three major and numerous smaller agricultural experiment stations scattered across the state. These facilities are conducting pure research and are staffed with scientists bent on adding to the existing knowledge base.

While county faculty members have received specific training in their respective fields and do participate in continuing education programs, it is impossible for them to be well informed across the whole gamut of agricultural disciplines. In essence, it would be highly unlikely to find someone who could answer many questions relating to forestry, entomology, plant pathology, agricultural economics, horticulture, and animal science.

It is anticipated that county faculty will reap intangible benefits as a byproduct of DDDI activities. In order to properly prepare samples for imaging, one will have to consider the diagnostic characteristics of the subject. For example, if an individual were preparing an insect specimen for diagnosis, that person must consider which features of the insect must be captured in an image to facilitate an accurate diagnosis. For identifying an insect, a diagnostician often must see a close-up shot of the mouthparts, the wing configuration, the body segmentation, etc. It is expected that through experience gained in preparing and submitting these samples and then communicating with the diagnostician and the client as to the diagnosis and treatment alternatives, county faculty will further develop their own diagnostic capabilities within other areas.

In addition, the positive reinforcement from clientele will encourage learning at the "teachable moment." In one instance, a county faculty member submitted a squash sample just weeks after receiving an imaging station. The sample was diagnosed within an hour of submission and a recommendation was immediately made. The timeliness of the recommendation saved an estimated 25% of the crop which would have been lost had the traditional sample delivery method been employed. This gave the faculty member an opportunity to make a substantive difference heretofore not possible—and a lesson not soon forgotten.

As county faculty members submit samples via DDDI, receive feedback and develop their own diagnostic skills, it is possible that the cyclical occurrences of pests, diseases, weeds, etc. may become evident. If these individuals could recognize (or even predict) and then respond to such trends, the environmental impacts and crop loss that are often associated with insect and disease outbreaks could potentially be held to a minimum. This is the high end of the predictive spectrum.

From an immediate standpoint, there are situations that, if addressed quickly, can be controlled to great advantage. For instance, if a pathogen that could cause an epidemic is identified early, DDDI provides the means by which professionals can be notified quickly as to potential problems. This could result in solving small problems that could become much bigger if not treated immediately. Such solutions result in fewer crop losses, less pesticide application, and much lighter environmental impact.

While the equipment that has been put in place with project funds was distributed with the understanding that diagnostic activities should always take precedence, county personnel were encouraged to use these new resources in other innovative ways. Many users have since embraced the technology and are using the equipment to augment and expand their activities in multiple program areas.

One lesson learned is that Distance Diagnostics is only one of many tools available for diagnosis. It is only as effective as the completeness of the information contained in the images and textual data. There are a number of circumstances where the diagnostician must ask for additional images and information. Occasionally physical samples must be requested in order to isolate biological processes beyond the capability of county faculty. In all cases, diagnosticians must make judgments based on their expertise and request additional information where appropriate.

New technologies and improvements to existing technologies are constantly changing the way we view DDDI. With the proliferation of mobile computing hardware and personal communications devices, for example, the possible development of portable imaging systems is becoming more realistic. These changes are not just taking place in the computing arena. Small, portable microscopes are now available that support digital photomicrography and are still capable of providing the same levels of magnification as their bench-top counterparts.

At this time, DDDI project members are investigating several opportunities for expansion of the project. Success of the core system has prompted numerous inquiries from other universities interested in developing similar systems. As a result, we have partnered with both Louisiana State University and the University of Illinois at Urbana-Champaign in designing and developing DDDI systems to enhance their respective diagnostic services. While working with these other institutions, in designing systems to meet the needs of their respective clientele, components are being developed that can be used in all DDDI systems to enhance the educational value of the information collected. Interest has also been expressed by other public-sector agencies and even some private-sector entities. Plans and proposals are being developed for addressing these needs. Additionally, International interest has generated requests for information and support worldwide.

As the DDDI system matures, we see more and more opportunities to enhance the usability and functionality of the system. When the system was first developed, it served its purpose well. Now, we are constructing features that will allow a diagnostician to forward samples to someone else for consultation or for initial diagnosis in a case where the primary diagnostician is unavailable for a timely diagnosis. We are also working towards providing both county faculty and diagnosticians with report generation capabilities as well. There are now security features in place to protect client information and we want to provide online resources for reference and research.

There are still many more functions and applications that can be built into and onto DDDI to further support the diagnostic efforts of our faculty. Among these, we are building a WWW-based library system that will allow images to be categorized and supplemented with descriptive text. This library will be available to the public and will eventually (when fully implemented) serve as a fully searchable visual diagnostic reference.

To date the Distance Diagnostics Through Digital Imaging System has exceeded expectations. There is abundant documented evidence of instances where DDDI has facilitated timely diagnosis or identification and intervention, preventing what could have potentially been individually (within a particular field) catastrophic crop or personal losses. The enthusiasm that has been demonstrated by county faculty using the system, as well as interest that has been expressed by other organizations in the system, has been unexpectedly great. Use of system components for public educational programs and presentations; for facilitating documentation of agricultural crop, herd or commodity quality; for identification of toxic plants consumed by livestock and humans; and for publicizing youth-program accomplishments have been visionary efforts made by county faculty. As system use expands and familiarity increases, ever more utility seems to become evident.

So far as payback is concerned, during initial implementation and since complete implementation of the system in July of 1999, greater savings from crop loss reduction has been documented than the total cost of implementing the system. However these cost savings do not include savings in cost of pesticide application and any necessary environmental remediation, nor savings from use of system equipment for immediate evaluation and recommendation, with no image/text submission by county faculty. Nor does it include the benefit of increased knowledge acquired by county faculty during evaluation of samples provided for DDDI analysis.

Additional system accomplishments and successes are yet to be seen. Continual growth of the system and its capabilities is occurring. Additional components are being developed internally and funding is stabilizing for a complete Internet Imaging System, of which DDDI is the main component.