Skip to main content
ARS Home » Northeast Area » Kearneysville, West Virginia » Appalachian Fruit Research Laboratory » Innovative Fruit Production, Improvement, and Protection » Research » Publications at this Location » Publication #343701
Title: Multi-modal sensor system for plant water stress assessment

Author
item KIM, JAMES - Monsanto Company
item Glenn, David

Submitted to: Computers and Electronics in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/10/2017
Publication Date: 7/11/2017
Citation: Kim, J.Y., Glenn, D.M. 2017. Multi-modal sensor system for plant water stress assessment. Computers and Electronics in Agriculture. 141:27-34.

Interpretive Summary: Evaluating and mitigating water stress is key to optimal productivity in crop production. A multi-modal sensing system was developed and evaluated for a rapid, automated scouting of plant stress status in an apple orchard with irrigated and dry trees. The system contained an infra-red thermometer array, a thermal imager, a multispectral camera, and two sets of Normalized Difference Vegetation Index (NDVI) sensors with a digital camera and an ultrasonic range finder. The experimental results identified significant differences of canopy temperature (2.6 _C) and NDVI (0.235) between the irrigated and dry trees, and supported further development of a low-cost real-time system for decision support of plant stress detection and management. The development of these sensors will increase crop yield and quality by indentifying stress on a small area or individual plant basis rather than a landscape scale.

Technical Abstract: Plant stress critically affects plant growth and causes significant loss of productivity and quality. When the plant is under water stress, it impedes photosynthesis and transpiration, resulting in changes in leaf color and temperature. Leaf discoloration in photosynthesis can be assessed by measurements of leaf reflectance changes, and leaf temperature changes in transpiration can be identified by thermography. To address these physiological properties, a multi-modal sensing system was developed and evaluated for a rapid, automated scouting of plant stress status in an apple orchard with irrigated and dry trees. The multimodal sensor system was installed on a mobile vehicle and includes an infrared (IR) thermometer array, a thermal imager, a multispectral camera, and two sets of Normalized Difference Vegetation Index (NDVI) sensors with a digital camera and an ultrasonic range finder. Soil water status was continuously monitored using soil moisture sensors that were installed at the 15-cm depth for both irrigated and dry trees. A low-cost solution of canopy temperature sensing was developed using an IR thermometer array and validated to substitute the thermal imager with the advantage of rapid 2-D thermal mapping at up to 10 Hz. An NDVI sensing system was developed to enhance the filtering process of the background noise signals by supplemental assistance from a digital camera and a range finder. NDVI responses and 2-D thermal maps were created while driving and recorded weekly for seven weeks during the growing season. The experimental results identified significant difference of canopy temperature (2.6_C) and NDVI (0.235) between the irrigated and dry trees and supported further development of low-cost real-time system for decision support of plant stress detection and management.