A tobacco field automatic identification method based on unmanned aerial vehicle image data

The invention discloses a tobacco field automatic identification method based on unmanned aerial vehicle image data. The identification method comprises the following steps of acquiring the unmanned aerial vehicle (UAV) image data at the initial stage of the seedling stage after the flue-cured tobacco is transplanted; carrying out the supervised classification and extracting a mixed type image oftobacco field film and buildings in the image data of the unmanned aerial vehicle; adopting the corrosion and dilatation algorithm to remove the building part in the mixed image to obtain a rough filmspatial distribution map of tobacco field; adopting the corrosion algorithm to remove noise broken spots from the rough tobacco field film spatial distribution map to obtain a fine tobacco field filmdistribution map; using a dilation algorithm to dilate the film distribution map of the fine tobacco field to obtain a preliminary extraction distribution map of the tobacco field; adopting a threshold segmentation method to r