The Great Infiltration

The common reed (Phragmites australis) is an aggressive threat to ecosystems across North America. Every state in the continental U.S. is affected by this invasive species, damaging wetlands and displacing native plants. Often seen along roadways in coastal areas, they can reduce both driver and boater travel visibility. As the common reed invades new environments, it decreases water flow and disrupts the soil quality, quickly reducing plant diversity and degrading a wetland’s natural ability to improve water quality and provide a flood control buffer. During droughts, dead reeds become highly flammable and increase wildfire risk in residential areas. They can also back up drainage ditches and stormwater systems, increasing flood risk. Plant and animal species are also negatively impacted by phragmites overgrowth. Large growths of tall reeds make it hard for animals to nest and limits access to food.

Tracking Down Invasive Plants

For many years, collecting invasive weed data has included ground and aerial surveys, manual mapping and recording, collecting samples, and public reporting (relying on community members to spot non-native species through observation). Some methods are more accurate than others, but each one requires significant time and resource commitments to do effectively. Today, AI-powered tools can make traditional methods significantly more effective.

Faster Detection, Identification, & Growth Prediction

AI provides an entirely new approach to the fight against invasive plants. With AI, we can use faster, more cost-effective, and comprehensive methods than traditional surveys alone:

• Easy identification. AI models can analyze drone imagery to recognize visual characteristics of phragmites and automatically generate precise boundary polygons of the invasive plant directly on images. This eliminates the need for specialized GIS expertise and the labor-intensive process of manually drawing polygons in mapping applications.
• Precision Treatment: After the scouting/data capture, drones can be calibrated to the collected data and dispense tiny droplets of chemical treatment approximately 10 feet above the reed stands. These targeted treatments reduce overall chemical exposure.
• Predicting growth: Models can be trained on aerial imagery and on-the-ground survey data to incorporate the effects of changing weather patterns (droughts, increased rainfall, etc.) to determine growth rates. By continuously updating models with new data, each year holds a greater opportunity to reduce phragmites spread.

Stopping it in its Tracks

PhragFinder, a machine learning model that quickly and accurately identifies phragmites in drone imagery, is now available on the BRYX platform, allowing environmental consultants to identify remediation areas. The model's output of geo-referenced polygons of the phragmites can be loaded into a sprayer drone for precise treatment. It’s now available in beta for free evaluation. Try PhragFinder today!