Evaluating NIR for precision agriculture grain harvest: design and development of a Grain Quality Monitor is the next logical step.

Precision Agriculture (altering agronomic inputs and farming processes such as tilling methods, irrigation or the application of fertilizer, herbicides and pesticides) has resulted in the introduction of new technology and sensors in farm equipment. Manufacturers have provided optional moisture and yield sensors on combines for over 10 years that have now become standard equipment on most new combines in North America. As a result of DOD navigational satellite availability and semiconductor technology advancement, GPS (Global Positioning System) has become standard for auto-steer and other functions on many tractors and combines.

For Precision Agriculture to realize its full potential, sensors that measure moisture, protein, oil, starch, fiber and other constituents in wheat, corn, soybeans and in other grains in real-time during harvest are required. These measurements, performed with a GQM (Grain Quality Monitor) coupled with a GPS, provide constituent field maps that are indispensable for effective precision agriculture.

Wheat producers in the northern Great Plains market grain under a quality payment system (protein premiums) that provides economic incentives for optimizing grain protein. In Montana, for example, the price spread between 13 percent and 15 percent protein Dark Northern Spring Wheat has averaged as much as $0.79 per bushel. Assuming this price spread, 1,000 acres of wheat yielding 50 bushels per acre with 15 percent protein would generate an additional $39,500 over that at 13 percent protein. In the Red River Valley, the wheat protein premium for the 2004 harvest reportedly averaged $70 per acre.

Protein concentration in grain is greatly influenced by the level of nitrogen (N) fertility, and there is significant spatial variation within a field. The usual uniform fertilizer application ignores this and results in over-fertilization in some areas and under-fertilization in others. Protein measurements provide a means of determining the deficiency or adequacy of soil N levels. Protein field maps created by a GQM/GPS make it possible to use variable-rate systems to compensate for the spatial variation in applying fertilizer. Conceptually, this optimizes grain yield and protein levels while reducing fertilizer and soil testing expenses.

The history of near infrared (NIR) begins in 1800 with the astronomer William Herschel and his work with telescopes. He demonstrated that there was light radiation beyond what we know as the visible spectrum. It wasn't until the early 1960s that Dr. Karl Norris in the USDA first applied it as an analytical technique for the prediction of protein and moisture in grain wheat. Since that time, because of huge advances in hardware and software technologies, NIR has become a widely used analytical tool and is the US Federal Grain Inspection Service standard for quantifying grain protein and moisture.

The first to use NIR for protein monitoring on combines was by Dr. Dan Long and Dr. Rick Engel of MSU (Montana State University) in 1996. A ProSpectra NIR analyzer was provided by Milestone Technology based on a concept proposed by Dr. David Mayes. This technology was sold to Textron and a joint venture with Case IH was established to commercialize it. The MSU researchers (Long and Engel) continued working with the ProSpectra until 2000 when Textron closed down their instrument division. In 2002 MSU began evaluating the Cropscan analyzer from NIR Technology Australia and used it through the 2004 harvest.

Dr. Long, now with the USDA in Pendleton, Oregon, also evaluated a Zeltex analyzer. Drs. Long and Engel are generally acknowledged to have more experience with real-time wheat protein monitoring than anyone in the world. They have created models using protein field maps for variable-rate fertilizers compensating for N variability within a field.

AgIR has been investigating NIR technology and instruments for more than a year. In September 2004, the company installed an evaluation Cropscan unit on a new John Deere 9860 combine at A.W.G. Farms, Inc. near Crookston, Minnesota. Harvesting and analyses were done by Gary Wagner, one of the farm's owners. Protein and moisture field maps were created from a few hundred acres of wheat. Samples were taken during the harvest and analyzed by laboratory instruments in local grain elevators. The combine measurements generally agreed within 0.2 percentage points of laboratory measurements (attached protein field map and protein value chart).

[ILLUSTRATION OMITTED]

To research the possibility of NIR's use in GQM for combines, AgIR assembled a strong core team of experienced high tech innovators, including Dr. Phil Williams, a pioneer in NIR. He was with the Canadian Grain Commission for over 35 years, where as Chief Scientist, he was instrumental in developing the use of NIR for protein and moisture measurements in wheat. He has experience with every NIR analyzer built and therefore is an invaluable guide in the design of an NIR system.

The two major parts to an analyzer system are a spectrograph and a grain sampler (see combine architecture and stationary architecture graphic). The Zeltex analyzer uses discrete NIR-emitting diodes and the Zeiss instrument uses an NIR diode array, while the ProSpectra and Cropscan are monochromatic spectrographs using dispersive elements and Silicon detectors. The ProSpectra sampling system collects reflected light from continuously moving grain. All other analyzers utilize either reflected or transmitted light from samples held fixed or transmitted light from moving samples.

[ILLUSTRATIONS OMITTED]

Sampling rates range from every 2 seconds up to 8 seconds depending on the system. A fiber optic cable pipe conveys the light to the spectrograph system in the combine cab for analyses. Data on protein and moisture are displayed in real-time during the harvesting and stored in a data base for later analyses.

Although a Grain Quality Monitor has not yet been designed with the features and the price/performance necessary to succeed in the Precision Agriculture market, AgIR has achieved a good understanding of the pros and cons of the various approaches taken by existing manufacturers. Technology is now available to design a GQM that meets the following specifications:

Easy to use: a simple, logical User Interface.

Integrates with the major combine systems or usable as a standalone unit.

Calibration is stable and transferable, stores multiple calibrations.

Rugged and reliable, designed for the severe environment.

Modular and portable, usable on or off the combine.

Serviceable remotely, wireless or internet access.

Priced below $10,000 initially and less with sales volume.

The NIR spectrum from 600 to 1550 nm (nano meters), depending on the particular instrument, contains "fingerprints" of several important constituents including protein and moisture that can be measured more accurately by NIR than with instruments currently sold with combines. The technology exists and the time is right for a very cost-effective, high price/performance GQM that is specifically designed and built for the combine market. This GQM will eliminate the requirement for a moisture sensor. It will be one that is modular and can be removed for use off the combine, one that is integrated with and utilizes existing combine systems, and one that is easily supported in the field.

Except for specific design patents on sampling systems and the Textron patent on reflected light from continuously moving samples, our research shows that most other claims are in the public domain and not patentable. NIR spectrometry has been used without restraint for many years. AgIR is currently evaluating this opportunity and considering strategic partners and investors.

Donald E. Hagge, is CEO of AgIR Company, Portland, Oregon. He can be reached at [e-mail].