Morshchavka S. Radioelectronic system of plant objects recognition

An object is the process of plant recognition on spectral coefficients. Aim - development of recognition methods for plants classification the field on useful and weed. Methods - remote sensing, mathematical statistics, statistical radiotechnics. Results - for the widespread types of the cultivated cultures and weeds the signature sets are experimentally got; possibility of classification is experimentally proved in the real-time mode with probabilities over 90%; possibility of automation of weed elimination on the plant-grower at the management by the executive aggregates of cultivators on results recognition is well-to-do. Novelty - the choice of signs for classification of separate plants as spectral coefficients of signals is grounded; it is first suggested to sensing the plants from lateral direction; it is offered in the tasks of the remote sensing for the receipt of high spatial permission to use the bistatic antennas system; the optimum methods of recognition for classification of separate plants got development; the classifier, allowing the fast parametric teaching and adaptation to changing properties of plants, is first offered. Level of introduction - model of classifier of plants developed in Zaporijja Institute of Construction and Technology of Agriculture Machines and during implementation of scientific projects in ZNTU. Sphere of the results use - constructors office of agricultural machines at development ecologically technologies for crop cultivation.