Method for carrying out real-time identification and targeted spraying on cotton field weeds

Abstract

The invention relates to a method for carrying out real-time identification and targeted spraying on cotton field weedsVisual system hardwares, comprising an additional light source, a CCD color video camera, an A/D converter, a DSP image processor, a D/A converter and an LCD (liquid crystal display) display; the additional light source can improve the situation in which light is not uniform or bright enough under natural conditions.The identification method comprises the following steps: a video camera installed at the front part of a weeding machine captures cotton field images in real time; after A/D conversion, the images are stored in a chip memory for being called by a DSP (Digital Signal Processor); the DSP extracts frame images from video streams and processes the frame images; green plants from soil backgrounds are segmented by the Cr characteristic method; the images are equally divided into four lines according to the arrangement of spray nozzles with each spray nozzle corresponding to a line; position features are used to recognize weeds among the lines; the weeds are recognized according to area and form features of blades; weeds position information is transmitted to a microcontroller; and the spray nozzles are opened to carry out accurate targeted spraying of weed killer to save pesticides and reduce pollution to the environment.

Description

technical field [0001] The invention belongs to the technical field of automatic control of agricultural machinery equipment, relates to an embedded machine vision system applied to weeding in cotton fields, and in particular to a method for real-time identification and spraying of weeds in cotton fields. Background technique [0002] Weed control in farmland is mainly based on the use of chemical herbicides. Now herbicides are mainly used for large-scale spraying. The herbicides sprayed on non-weed areas not only cause waste, but also cause serious air pollution, water pollution, soil pollution, and agricultural product pollution. In addition, agricultural production The labor force is also constantly decreasing. In order to reduce the harm of weeds, reduce environmental pollution, improve the utilization rate of pesticides, liberate the labor force of weeding, and realize the automation and scientificization of weed control is worth exploring. [0003] Field weeds are the...

AI Technical Summary

Problems solved by technology

There are about 1,500 kinds of weeds in farmland in my country, including 38 kinds of malignant weeds, 94 kinds of regional weeds, and 364 kinds of common weeds; the harm of field weeds to crops mainly includes: competing with crops for water, Fertilizers and glare will affect the yield and quality of crops; increase management labor and production costs, and weeds will also reduce the quality of agricultural products; in addition, weeds will also promote the breeding and spread of diseases and insect pests; Over 2.39 million tons of grain, over 20,000 tons of cotton, nearly 200,000 tons of oil crops, and over 1.37 million tons of other crops have been lost due to weed damage; weed damage alone causes an annual loss of about US$76.3 billion on 1.5 billion hectares of arable land worldwide

[0004] At present, the most important weeding method in the world is chemical weeding. In developed countries, the amount of herbicides used accounts for nearly half of the amount of pesticides use

Images