A research team from the University of Arkansas Division of Agriculture developed a method of measuring the ‘greenness’ of corn leaves using a digital camera and commercially-available software. Funding for the project was provided by the University of Arkansas Agricultural Experiment Station. The research appears in the September–October 2011 issue of Crop Science. This research was also presented in 2009 at the Crop Science Society of America’s annual meeting in Pittsburg, PN.
In greenhouse experiments, corn was grown over a wide range of soil-N treatments, from 0 to 840 mg N per 3-L pot. The upper-most collared leaf from 3- to 5-leaf plants was removed and photographed under fluorescent lighting using three different digital cameras. SPAD measurements were also made on each leaf, and then leaves were dried and analyzed for total N. From a field experiment the upper-most leaf was removed from 5-leaf plants and photographed indoors under fluorescent and incandescent lighting and outdoors under full sun and shaded conditions. Included in the photographs from the greenhouse and field experiments were standard disks of known color that served as internal standards; one disk was light yellow and one disk was dark green.
Digital images were processed using SigmaScan Pro 5 and a macro that combines the hue, saturation, and brightness values of each image into one composite number, the Dark Green Color Index (DGCI). The user selects the portion of the color spectrum that the software recognizes, in this case from light yellow to dark green, and the software determines the hue, saturation, and brightness in each image. The macro allows batch analysis of images and calculates DGCI. The known DGCI values of the standard disks are used to correct for differences in color sensitivity among cameras and lighting conditions.
Without using a correction from the internal standards, there were large differences in DGCI values among cameras and among lighting conditions. By using the color disks as internal standards, nearly all of the differences in DGCI among cameras were eliminated. Likewise, the internal standards greatly improved the agreement in DGCI values among lighting conditions (Panels A, B, C, and D). Corrected DGCI values agreed closely with SPAD values (r2 = 0.91) and with leaf N concentration (r2 values ranged from 0.81 to 0.89). This research agrees well with field research reported in the March-April issue of Agronomy Journal that also found a close agreement between DGCI and leaf nitrogen concentration. The Agronomy Journal article also showed that DGCI values at tasseling in corn were closely associated with corn grain yield.
Larry Purcell, who is the corresponding author of the research, said “This research developed techniques that allow a digital camera or smart phone to be used as a diagnostic tool for leaf nitrogen. The inclusion of color disks as internal standards opens up the possibility of taking digital images directly in the field and uploading the images to a server for immediate analysis of plant nitrogen status.”
Continued research at the University of Arkansas is focused on calibrating the amount of nitrogen to apply to corn to recover yield potential based upon DGCI measurements made at V8. Research is also developing DGCI measurements as a diagnostic tool for other non-leguminous crops.
Robert L. Rorie, Larry C. Purcell, Douglas E. Karcher and C. Andy King
The Assessment of Leaf Nitrogen in Corn from Digital Images
Crop Science 2011 51:2174-2180