Multispectral vision based efficient weeding robot

Abstract

The invention discloses a multispectral vision based efficient weeding robot comprising a travelling car, a transverse connecting device, a weeding cutting device and a master control system, wherein the transverse connecting device, the weeding cutting device and the master control system are arranged on the travelling car and are connected with each other. The multispectral vision based efficient weeding robot is capable of rapidly and effectively recognizing farmland weeds, strong in real-time property and suitable for actual field work; the weed clearing works inside and outside a row are simultaneously finished through the combination of a simple cutting structure and a pesticide spraying way, not only is the load of the travelling car reduced, but also the weeds can be thoroughly cleared, the weeding efficiency is high, and the problem for clearing the weeds inside and outside the row is effectively solved; and due to the adoption of a derusting device, the work efficiency of a cutting tool is increased effectively, the service life of the cutting tool is prolonged, and the comprehensive performance of the weeding robot is improved.

Description

technical field [0001] The invention relates to a high-efficiency weeding robot based on multispectral vision, in particular to a robot capable of using multispectral images to identify field weeds and perform rapid weeding, belonging to the field of agricultural machinery applications. Background technique [0002] Farmland weeds are the enemy of agricultural production. They will affect the growth and development of crops, reduce the yield and quality of crops, and the harm to crops cannot be ignored. my country spends a lot of manpower, material resources and financial resources to carry out weed cleaning operations every year. Therefore, researching an efficient weeding robot has important practical significance for improving agricultural production efficiency and grain and other crop yields. [0003] At present, the main weeding methods are: manual weeding, mechanical weeding and chemical weeding. Among them, the low efficiency of manual weeding has been gradually eli...

AI Technical Summary

Problems solved by technology

However, in the actual field environment, since the reflection of green leaves to light is affected by many factors, the leaf features extracted in visible light cannot well represent the difference between crops and weeds.

[0006] In addition, in many devices that use cutting and weeding, the knives are easily contaminated with plant tissue fluid or water and other compounds when cutting weeds and rust, resulting in a short life of the knives and increasing the cost of use, which is not conducive to popularization.

Images