Assessing the Capability and Potential of LiDAR for Weed Detection

Area of Science:

• Agricultural Engineering
• Remote Sensing
• Agronomy

Background:

• Conventional uniform field spraying incurs high herbicide costs and environmental impact.
• Site-specific weed management (SSWM) requires accurate weed localization.
• Technological advancements enable automated weed detection using drones and ground-based sensors.

Purpose of the Study:

• To assess the capability of Light Detection and Ranging (LiDAR) sensors for detecting and locating weeds.
• To determine the detection limits of LiDAR technology for artificial weed targets.

Main Methods:

• Two trials utilized artificial targets of varying sizes and heights to establish LiDAR detection limits.
• A third trial involved scanning a wheat plot with diverse weed species at different heights above the crop canopy.
• Stationary LiDAR was employed to evaluate real-world field detection capabilities.

Main Results:

• LiDAR detectability of artificial targets was influenced by target size and orientation relative to the sensor.
• Detection accuracy varied with scanning distance.
• In a field trial, LiDAR successfully detected 100% of weeds based on height differences from the wheat canopy.

Conclusions:

• LiDAR technology demonstrates high potential for accurate weed detection in agricultural settings.
• Height differentiation between weeds and crops is a key factor for successful LiDAR-based weed mapping.
• LiDAR offers a viable solution for enhancing site-specific weed management strategies.