Ground-Based Remote Sensor Development  Plant Health Determination

Leaf discoloration often signifies a plant health problem.  Discoloration may be associated with nutrient deficiency or toxicity, disease, or insect infestation.  Remote sensing allows measurement of reflected light energy in one or more wavebands.  In many cases, spectral information gained through remote sensing may be useful for identification and diagnosis of plant health problems as they develop.

There are several sources of spectral data.  Satellite-based systems gather data over large areas, and have the advantage of being completely automated (from the end-user's perspective).  Disadvantages of satellite-based systems include:  (1) they are weather-dependent, since data cannot be gathered when clouds obscure the target; (2) resolution is limited; and (3) the end-user has no control over data collection timing or procedures.  Airplane-based systems overcome most of the disadvantages of satellite-based systems, however cost of operating planes often prohibits their use.  Un-manned aircraft may provide a viable solution, but cost remains a significant limitation.  Sensors and Controls Lab personnel have focused remote-sensing efforts on a ground-based system designed to mount on existing field equipment.

Sensors and Controls Lab research has focused on using spectral information to diagnose nitrogen deficiency in cotton.  The most current prototype includes an illumination source, a measurement system, and a processing unit.  The illumination source and measurement system include features that make reflectance quantification independent of ambient lighting conditions.  The processing unit implements a feed-forward neural network which uses the spectral data to assess cotton plant nitrogen status.  Variations of the system have been field-tested during two growing seasons.  In the most recent set of tests, conducted in west Tennessee during 2001, the system accurately classified cotton plant nitrogen status at the pinhead square growth stage as high, medium, or low with 90% accuracy.  Additional improvements are expected as research continues.

Sui_graph.jpg (6,156 bytes)

Sui_w_prototype1.jpg (11,404 bytes)

Chris_w_prototype2.jpg (11,540 bytes)
Publications:

Sui, R.  1999.  A ground-based real-time remote sensing system for diagnosing nitrogen status in cotton plants.  Ph.D. dissertation, The University of Tennessee, Knoxville, Tenn.

Whitten, C.L.  2002.  Development of a ground-based remote sensing system with modulated illumination for diagnosing nitrogen status in cotton.  M.S. thesis, The University of Tennessee, Knoxville, Tenn.

Industry Partner:
Cotton Incorporated
For more information about this project, please contact:

John Wilkerson, Ph.D.
The University of Tennessee

2506 E. J. Chapman Drive
Knoxville, TN  37996-4531
PHONE:  (865) 974-7266

wilkerj@utk.edu