Original Source: The Cool Down
A tool for … experts to assess plant health’
Traditional tree health monitoring has been a hands-on, labor-intensive job.
Scientists in Japan have developed an artificial intelligence-driven “Plant Doctor” that monitors and diagnoses tree health in cities. Without ever touching a leaf, the groundbreaking system keeps urban green spaces thriving. The technology could revolutionize how cities care for their tree populations in a more efficient and less invasive way.
Urban trees are more than pretty scenery and are essential for healthy cities. They provide the clean air we breathe and cool hot city streets, reducing the urban heat island effect. Trees also provide shade, cool recreational areas, and even increase property values.
City trees, unfortunately, face many threats, including pests, diseases, and a changing climate. Traditional tree health monitoring has been a hands-on, labor-intensive job. Botanical experts have to physically inspect trees, a costly and time-consuming process.
Recognizing this challenge, a joint research team from Waseda University and Ryukoku University developed Plant Doctor. Their study in the journal Measurement mentions that the AI system uses video cameras to assess tree health.
The first author, Marc Josep Montagut Marques, explained the innovation behind the hybrid system: “Machine-vision techniques such as segmentation have great applications in the medical field. We wanted to extrapolate this technology to other areas, such as plant health.”
The system combines several machine-vision algorithms. First, YOLOv8 and DeepSORT work together to identify and track individual leaves in video footage. Only the highest quality images of each leaf are selected.
Then, the DeepLabV3Plus algorithm performs detailed image segmentation. It gives precise measurements of any damage on the leaf, such as spots caused by bacteria, fungi, or pests.
The Plant Doctor AI offers huge benefits. It has the cost-effective capacity to monitor vast numbers of trees. Using existing city infrastructure, like maintenance vehicle cameras, allows for large data collection and mapping. The non-invasive processing means there is no need to take samples or damage the trees for a checkup.
“We have provided a tool for botanical experts to assess plant health in one solution,” Marques remarked. The team performed a successful system validation in Tokyo with accurate health diagnoses.
Health and location information gave insights into individual plant issues and disease patterns across urban areas. This data leads to strategic interventions rather than routine, time-consuming monitoring.