FIELD: computer technology.
SUBSTANCE: method for the detection of a weed type in natural environment contains video capture of a digital weed image among cultivated crops in natural environment; contouring in the digital image of areas with a pre-determined characteristic of color and structure in RGB-color model to form at least one contoured area containing pixels related to a weed withing a boundary contour; transmission of the digital image for further research including the application to the captured digital image of at least one trained correlation function of a classifier, including neural network classifiers and object-oriented identifiers both for weed and separate leaves, to detect the weed, wherein the correlation function has an access to names of weed types together with a set of metadata for each type of the weed, and metadata is used as input values for a neural network system generating a weed name together with probability for one or several weed types; and reception of a weed name for the weed of the captured image and a value of probability of coincidence between the weed name and the weed of the captured digital image.
EFFECT: detection of a weed name with high probability.
18 cl, 8 dwg
Title | Year | Author | Number |
---|---|---|---|
WEEDS IDENTIFICATION IN NATURAL ENVIRONMENT | 2017 |
|
RU2735151C2 |
LASER SURGICAL SYSTEM AND METHOD FOR USE THEREOF FOR EVAPORATION OF LIVING TISSUE | 2015 |
|
RU2582213C1 |
METHODS AND DEVICES FOR ANALYTICAL MEASUREMENT | 2018 |
|
RU2777489C2 |
DETECTION OF PLANTS DISEASES BY CLASSIFICATION OF PLANTS PHOTOGRAPHY | 2019 |
|
RU2805670C2 |
GRAIN FIELD YIELD PREDICTION | 2017 |
|
RU2769288C2 |
METHOD AND SYSTEM FOR IMAGE COLOUR DEFINITION | 2011 |
|
RU2567500C2 |
METHOD OF APPLYING A SPRAYING AGENT ON A FIELD | 2018 |
|
RU2767347C2 |
ANCHORS FOR NAVIGATION BASED ON LOCATION AND APPLICATION OF ADDITIONAL REALITY | 2013 |
|
RU2639672C2 |
SYSTEM AND METHOD FOR CAPTURING MEASURING IMAGES OF MEASURED OBJECT | 2017 |
|
RU2758914C2 |
METHOD OF RECORDING STATE OF PUMP UNIT | 2017 |
|
RU2689271C1 |
Authors
Dates
2022-01-21—Published
2017-05-05—Filed