Title : Research progress on laser weeding robots
Abstract:
The extensive use of chemical herbicides has led to deteriorated soil environments, reduced biodiversity, and pesticide residues that endanger human health and damage ecological systems, thereby compromising the sustainable development of agriculture. The cultivation of Chinese medicinal herbs, premium vegetables, and green organic products is subject to strict restrictions on herbicide usage, leaving manual weeding as virtually the only available option. Laser weeding represents a highly efficient, non-contact, and precise physical weed control method. It exhibits strong adaptability to various agronomic practices, high weeding efficiency, low seedling damage rates, and minimal disturbance to soil and crops, making it well-suited to China's diverse agricultural conditions characterized by multiple crop species, complex planting patterns, and varied terrains and soil types. This research has made breakthroughs in several key technologies, including lightweight high-precision AI recognition of weeds, three-dimensional precise localization of meristematic tissues (shoot apices), sub-millimeter dynamic laser beam target tracking, collaborative intelligent planning and adaptive regulation of multi-laser beams, and multi-source fusion navigation with range-extended wire-controlled chassis. These advancements have enabled high-precision identification and efficient dynamic targeted elimination of small-target weeds in complex environments, supporting round-the-clock day-and-night operations. The equipment is primarily applicable to high-value cash crops, Chinese medicinal herbs, and grain and oil crops grown in dryland conditions, with optimal weeding performance achieved when weeds are at the 2–4 leaf stage. Validation results demonstrate that weed recognition accuracy exceeds 95%, positioning precision reaches ±0.2 mm, tracking precision achieves ±0.3 mm, and the average elimination efficiency is 0.3 s per plant. The weeding throughput reaches 87,000 plants per hour, surpassing manual labor by more than 50 times, with a weed removal rate of ≥90% and a seedling damage rate of ≤4.5%. This technology provides a novel technical approach for efficient and green weed prevention and control, and offers equipment support for the reduction of chemical pesticides and the sustainable development of agriculture.
