Identification of the Spectral Patterns of Cultivated Plants and Weeds: Hyperspectral Vegetation Indices

The accurate recognition of weeds on crops supports the spot application herbicides, high economic effect and reduction pesticide pressure agrocenoses. We consider approach based quantitative spectral characteristics plant objects to be most appropriate for development methods herbicides. made test plots with different species composition cultivated weed plants experimental fields scientific crop rotation Federal Research Center Biological Plant Protection. These form basis agrocenoses Krasnodar Krai. Our primary subjects are sunflower (Helianthus annuus L.), corn (Zea mais L.) soybean (Glycine max (L.)). Besides plots, pure mixed backgrounds were identified, represented by following species: ragweed (Ambrosia artemisiifolia California-bur (Xanthium strumarium red-root amaranth (Amaranthus retroflexus white marrow (C. album field milk thistle (Sonchus arvensis L.). used Ocean Optics Maya 2000-Pro automated spectrometer conduct high-precision ground-based spectrometric measurements selected plants. calculated values 15 generally accepted index dependencies data processing from ground hyperspectral They aided in evaluating certain vegetation parameters. Factor analysis determined relationship structure variable indices into individual factor patterns. variance assessed information content indicators within limits factors. concluded that under consideration characterized homogeneity signs according make up However, cases, it is possible identify backgrounds, both generalized factorial coefficients. research results important validation remote aerospace observations using multispectral instruments.