Weed management during peanut cultivation is one of the common challenges for farmers. Particularly harmful to crops is Johnson Grass (or Sorghum halepense) — the most aggressive weed in the world, which outcompetes cultivated plants, actively absorbs moisture and nutrients.
Timely treatment of small weeds, including with herbicides, is crucial for plant health and vitality. The most resilient technological solution is the use of drones combined with professional software.
As a result, one company in Argentina was able to optimize weed control of Johnson Grass on peanut crops by using RGB drones (working in the visible spectrum), RTK (Real-Time Kinematic) technology, and PIX4Dfields software. This technological approach enabled rapid detection of pests, creation of maps for targeted "green-on-green" spot spraying, and cost reduction.
Johnson grass (Sorghum halepense)
Project details
Location: Cordoba, Argentina
Project duration: 5 hours
Area surveyed: 1500 hectares
Software used: PIX4Dfields
Hardware used: RGB drones
Images captured: Total images: 20,000 RGB
Processing time: 10 minutes per 100 hectares (on average)
GSD: 2cm
PIX4Dfields tools
Orthomosaic of peanut field processed in 5 minutes in PIX4Dfields
Within 5 minutes, using PIX4Dfields' fast and offline processing, the team created an RGB orthomosaic of the field.
The key tool used was the Magic Tool — a PIX4Dfields solution based on AI, which automatically detects weeds and other anomalies in real-time from RGB drone images. As a result, herbicide application becomes targeted and affects only the weed plants.
Key steps included
Weed selection with the Magic Tool in PIX4Dfields
-
AI Training and Detection: To accurately detect weeds, 5 points with weeds and 5 crop areas (peanuts) were selected. Based on these examples, the AI tool was trained to differentiate weeds from crops.
-
Operation Layer Creation: The final detection layer obtained using the Magic Tool was saved as an operation layer. It can be used for targeted herbicide application in the field, enabling the quick creation of a task map for machinery.
Thanks to the new Magic Tool version (designed to minimize user intervention), this workflow precisely identified Johnson grass, even in fully developed peanut crops.
Wanted (weeds) and unwanted (peanut crop) areas of the field marked with PIX4Dfields’ Magic Tool
The targeted marking of weeds allows for precise herbicide doses and identification of target areas for application. This results in a 93% savings in chemical inputs.
Magic Tool has detected all weeds on the field in PIX4Dfields
Limiting herbicide application exclusively to weed control reduces environmental impact, preserves crops, and saves costs by avoiding treatment of non-target areas of the field.
Prescription map generation and herbicide dosage assigned in PIX4Dfields
Total savings and benefits
Table showing total savings from a single targeted herbicide application versus blanket application on a 45-hectare peanut crop field
- Time and cost savings: with fast processing and the Magic Tool, weed identification and prescription creation in PIX4Dfields were completed in just a few minutes. With the targeted application 93% reduction in input usage (herbicide and water) was achieved.
- High detection accuracy: combining RTK drones with 2 cm/px orthomosaics captures even the smallest details.
- Improved sustainability: avoiding widespread herbicide spraying reduces input usage and ecosystem impact.
- Easy integration: the Operation layer created with the Magic Tool is easily exported - no third-party software needed - for use in variable-rate spraying equipment (spray drones, tractors, field sprayers).
Overall, the entire project took 5 hours, thanks to the seamless integration of AI technology into agriculture. Weed control and crop preservation became more efficient and environmentally friendly.
PIX4Dfields helps farmers worldwide optimize production processes and use agrochemicals responsibly.
