Remote sensing technology has been developed and applied to provide spatiotemporal information on crop stress for
precision management. A series of multispectral images over a field planted cotton, corn and soybean were obtained by a
Geospatial Systems MS4100 camera mounted on an Air Tractor 402B airplane equipped with Camera Link in a Magma
converter box triggered by Terraverde Dragonfly® flight navigation and imaging control software. The field crops were
intentionally stressed by applying glyphosate herbicide via aircraft and allowing it to drift near-field. Aerial multispectral
images in the visible and near-infrared bands were manipulated to produce vegetation indices, which were used to
quantify the onset of herbicide induced crop stress. The vegetation indices normalized difference vegetation index
(NDVI) and soil adjusted vegetation index (SAVI) showed the ability to monitor crop response to herbicide-induced
injury by revealing stress at different phenological stages. Two other fields were managed with irrigated versus nonirrigated
treatments, and those fields were imaged with both the multispectral system and an Electrophysics PV-320T
thermal imaging camera on board an Air Tractor 402B aircraft. Thermal imagery indicated water stress due to deficits in
soil moisture, and a proposed method of determining crop cover percentage using thermal imagery was compared with a
multispectral imaging method. Development of an image fusion scheme may be necessary to provide synergy and
improve overall water stress detection ability.
Aircraft routinely used for agricultural spray application are finding utility for remote sensing. Data obtained from remote sensing can be used for prescription application of pesticides, fertilizers, cotton growth regulators, and water (the latter with the assistance of hyperspectral indices and thermal imaging). Digital video was used to detect weeds in early cotton, and preliminary data were obtained to see if nitrogen status could be detected in early soybeans. Weeds were differentiable from early cotton at very low altitudes (65-m), with the aid of supervised classification algorithms in the ENVI image analysis software. The camera was flown at very low altitude for acceptable pixel resolution. Nitrogen status was not detectable by statistical analysis of digital numbers (DNs) obtained from images, but soybean cultivar differences were statistically discernable (F=26, p=0.01). Spectroradiometer data are being analyzed to identify narrow spectral bands that might aid in selecting camera filters for determination of plant nitrogen status. Multiple camera configurations are proposed to allow vegetative indices to be developed more readily. Both remotely sensed field images and ground data are to be used for decision-making in a proposed variable-rate application system for agricultural aircraft. For this system, prescriptions generated from digital imagery and data will be coupled with GPS-based swath guidance and programmable flow control.