A Real-Time Weed Mapping and Precision Herbicide Spraying System for Row Crops

Area of Science:

• Agricultural Engineering
• Precision Agriculture
• Computer Vision

Background:

• Traditional weed management relies on uniform herbicide application, leading to overuse and environmental concerns.
• Automated systems are needed to optimize herbicide use and improve crop yield in row crops.
• Existing machine vision algorithms for weed detection face challenges in accuracy and processing speed.

Purpose of the Study:

• To develop and field-test an automated weed mapping and variable-rate herbicide spraying (VRHS) system for row crops.
• To enhance weed detection accuracy and real-time herbicide application capabilities.
• To evaluate the system's performance in practical weed management scenarios.

Main Methods:

• Developed a machine vision sub-system utilizing a custom threshold segmentation method and an improved particle swarm optimum (IPSO) algorithm for image segmentation.
• Implemented a lateral histogram-based algorithm for rapid weed map extraction to determine real-time herbicide application rates.
• Integrated a central processor with high logic operation capacity and a custom monitoring system for real-time visualization and performance evaluation.

Main Results:

• The IPSO algorithm successfully segmented weeds in corn crops with a reduced error rate of 0.1% (compared to 7.1% for traditional PSO).
• Achieved a high processing speed of 0.026 s/frame with the IPSO algorithm.
• Demonstrated a weed detection to chemical actuation response time of 1.562 seconds for the integrated system.

Conclusions:

• The developed VRHS system effectively meets real-time data processing and actuation requirements for practical weed management.
• The IPSO algorithm significantly enhances weed detection accuracy and processing speed in agricultural applications.
• This automated system offers a promising solution for optimizing herbicide use and improving the efficiency of row crop farming.