/AnMtgsAbsts2009.53536 Characterizing Leaf N with Digital Images in Corn and the Association of "Greeness" with Yield.

Monday, November 2, 2009
Convention Center, Exhibit Hall BC, Second Floor

Robert Rorie, Larry Purcell, Douglas Karcher, Andy King and Morteza Mozaffari, Univ. of Arkansas, Fayetteville, AR
Poster Presentation
  • ASA poster 2009.pdf (2.1 MB)
  • Abstract:
    The environmental implications of nitrates coupled with the growing usage and cost of N fertilizers have compelled agronomist to develop quick and accurate methods of determining plant N. Our objective was to use a digital camera and image-analysis software to quantify the “greenness” of corn (Zea mays L.) leaves that could serve as a relative indicator of plant N status. In addition to Digital Color Analysis (DCA) of leaves, we also evaluated relationships among SPAD, total leaf N, and the use of internal standards for camera calibration. Field experiments were conducted in Fayetteville, Marianna, Keiser, and Rohwer, Arkansas over a wide range of N treatments (0 to 340 kg N/ha). At tassleing, we measured leaf greenness (DCA), SPAD and total N on the uppermost collared leaf. “Greenness” was quantified as a dark green color index (DGCI), with values from 0 to 1, from hue, saturation, and brightness values as determined by the software. Digital images of leaves from field experiments were initially taken using a black cloth as a background, but because of background noise and light reflectance, SPAD and DGCI agreement was only moderate (average r2 values = .54). In greenhouse experiments, digital images of leaves were taken against a pink background, and the association between DGCI and SPAD values was much closer (average r2 values =.92). Both DGCI and SPAD showed close, positive linear relationships with leaf N concentration when leaf N was less than 29 mg N/g, but above this concentration the relationships became curvilinear and reached a plateau. In greenhouse experiments, we were able to eliminate differences in DGCI values from different cameras by including disks of known DGCI values in each image. These data indicate a promising role for using digital images to quantify in-season N status of crops.