Spectral reflectance at sub-leaf scale including the spatial distribution discriminating NPK stress characteristics in barley using multiway partial least squares regression
Abstract:This paper introduces a methodology able to discriminate between non-stressed plants and N, P and K stress symptoms in spring barley grown under controlled conditions, utilizing the spectral and spatial dimensions simultaneously. Nine spectral measurements in the range 450-1000 nm were taken for each plant. The measuring points were spatially located at the tip, middle and base of the last three fully developed leaves. This design generated a four-way data set consisting of measurements as a function of (i) the specific plant, (ii) the spectral wavelength, (iii) the plant leaf position and (iv) the position on the leaf. Multiway partial least squares regression analysis with dummy variables was able correctly to classify the four nutrient conditions with 92% accuracy regardless of the respective growth stages within a time window of 2 weeks. The addition of the spatial dimension to the spectral dimension proved to be a promising nutrient diagnostic tool. Without performance loss it was possible to reduce the hyperspectral resolution to a resolution of three wavelengths. The three selected 2 nm wide bands were R450, R700 and R810, which agrees well with the literature on plant spectral reflectance in relation to nutritional stress.
Document Type: Research Article
Affiliations: 1: Ministry of Food, Agriculture and Fisheries, Danish Institute of Agricultural Sciences, Department of Agricultural Engineering, Research Centre Bygholm, DK-8700 Horsens, Denmark 2: Nordic Genebank, SE-23053 Alnarp, Sweden 3: Chemometrics Group, Food Technology, Department of Dairy and Food Science, The Royal Veterinary and Agricultural University, DK-1958 Frederiksberg C, Denmark
Publication date: January 1, 2007