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

Abstract

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 remove impurity plots from the preliminary extraction tobacco field distribution map to obtain a tobacco field spatial distribution map, wherein the impurity plots are plotsother than tobacco fields. The automatic recognition of the tobacco field image is realized and the distribution condition and the planting area of the tobacco fields are analzyed, the tobacco field area data can be monitored in real time, and the precision and efficiency of the tobacco field spatial distribution and area investigation can be improved.

Description

technical field [0001] The invention relates to the field of tobacco field monitoring, in particular to an automatic identification method for tobacco fields based on drone data. Background technique [0002] my country is one of the most important tobacco production and consumption countries in the world. Flue-cured tobacco is the type of tobacco with the largest cultivation area in my country and the world. It is the main raw material of the cigarette industry, an important agricultural product in my country's economic crops, and one of the bulk agricultural products exported by my country. Tobacco planting information is an important basis for tobacco production management, can provide decision support for tobacco planting and acquisition control, and is an important economic information for national macro-management and decision-making. The estimation of tobacco planting area is helpful to grasp the situation of tobacco planting and supervise the planting scale of tobac...

AI Technical Summary

Problems solved by technology

[0004] Compared with bulk crops, tobacco plots are fragmented and scattered, and there are many crops planted in the same period, which are easy to be mixed with other crop types. For remote sensing data with spatial resolution and time phase, radar data usually have low resolution and cannot meet the needs of tobacco. The need for real-time monitoring of spatial distribution

Images