Premium
Classification of Land Use and Crop Maturity, Types, and Disease Status by Remote Reflectance Measurements 1
Author(s) -
Tinker R. W.,
Brach E. J.,
LaCroix L. J.,
Mack A. R.,
Poushinsky G.
Publication year - 1979
Publication title -
agronomy journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.752
H-Index - 131
eISSN - 1435-0645
pISSN - 0002-1962
DOI - 10.2134/agronj1979.00021962007100060025x
Subject(s) - hordeum vulgare , crop , avena , agronomy , rapeseed , growing season , vicia sativa , mathematics , environmental science , remote sensing , biology , poaceae , geography
Remote sensing techniques supply useful crop information in large area crop survey programs. In such programs, however, the spectral and spatial resolution, measurement geometry, and environmental correction cap abilities have been limited by the fixed and truly remote nature of the sensor platform. To extend the applications of existing systems, or to use new systems, some workers have started small‐scale, ground‐based programs, such as the one described in this paper. The purpose of the experiment was to classify land use categories, crop maturity, type, and disease status by discriminant analysis of high resolution, normalized visible (350 to 750 nm) and infrared (750 to 1,850 nm) spectral reflectance data which were acquired, at an oblique angle, from field plots of sod, soil, and various cereal and broadleaf crops throughout the 1976 growing season. Biological data (plant heights, growth stages, leaf chlorophyll) were acquired concurrently with plot spectral measurements so that the results of the classifications could be explained in term of plot development or condition. Land use categories (sod, soil, crop), stages of crop maturity (vegetative, flowered), crop type [rapeseed ( Brassica napus L. and B. campestris L.), fababean ( Vicia faba L.), soybean ( Glycine max L. Merr.), and cereals], and disease status (diseased or healthy cereals) were classified from the spectral data with 95.2, 93.3, 90.7, and 67.9% accuracy, respectively. The accuracy of classification for specific cereals was low [wheat ( Triticum aestivum L, T. durum L.), 53.5%; oats ( Avena sativa L), 67.7%; barley ( Hordeum vulgare L.), 50.0%]. Partially due to stress, the growth habit and, therefore, reflectance of these crops was similar. The results of the spectral classifications were supported by trends observed in plot development or condition. Thus, it was possible to discern land use categories, general crop maturity and specific or general crop types from the analysis of remotely acquired reflectance data. Such parameters would supply useful information in a functional crop reporting system.