Drone-based control of invasive apple snails — NARO develops labor-saving system for spot application of chemicals

The Smart Lowland Crop Rotation Group within the Division of Crop Rotation Research for Lowland Farming, Kyushu Okinawa Agricultural Research Center, National Agriculture and Food Research Organization (NARO), which includes Group Leader Kimiyasu Takahashi and Senior Researcher Senlin Guan, announced on October 31 that they have developed a control system that automatically creates damage prediction maps for paddy rice damage caused by Pomacea canaliculata (commonly known as invasive apple snails) through drone imaging and enables spot application of chemicals. This is expected to contribute to labor savings and cost reduction in paddy rice cultivation. The results were published in Agriculture.

P. canaliculata is a freshwater snail that is a designated invasive alien species and harmful animal or plant. They are known to become active when water temperatures exceed 17℃ and preferentially feed on small seedlings about two weeks after transplanting. It is also known that damage is more likely to occur under conditions where water depth exceeds 4 centimeters.

To address this, various methods have been combined, including maintaining shallow water depth of approximately 1 centimeter in paddy fields (shallow water management) and chemical application. However, in recent years, damage has occurred due to flooding even with shallow water management because of extreme weather changes. There are also concerns that climate change is causing the expansion of habitats, an increase in population due to warm winters, and an increase in overwintering individuals. Generally, paddy fields have elevation differences, and traditionally farmers performed work to level the soil height, but this has become difficult due to the declining agricultural workforce. Although chemical application is effective, high costs related to applying it across entire fields has also been an issue.

The research group developed an efficient control method using drones. First, to apply chemicals only where needed, they create a field elevation difference map from drone aerial images. Next, using a newly developed damage prediction algorithm (patent obtained), they converted this map into a "damage prediction map" that identifies areas with water depths exceeding 4 centimeters. Finally, this map is input into a drone capable of variable-rate application and chemicals are applied only to areas where damage is predicted (spot application).

In verification tests, aerial photography was conducted after plowing in farmers' fields in Saga Prefecture, the damage prediction map information was input into a drone capable of variable-rate application, and spot application was performed after rice transplanting in late June.

As a result, chemical usage was reduced by 45% compared with full-field application. It was confirmed that the damaged area could be suppressed to 10%, similar to full-field application.

Furthermore, to promote wider adoption, the researchers developed a web user interface that can automatically create elevation difference maps and damage prediction maps from drone images without requiring specialized knowledge. Each map can be easily created by uploading images and selecting the drone model, among other simple steps. User testing is currently underway, and next fiscal year they plan to expand testing to large-scale farmers in cooperation with regional public testing stations.

Takahashi commented: "Drone-based analysis that only specialists could perform can now be done automatically by farmers themselves through a simple interface. We expect this will help promote smart agriculture and data-driven agriculture."

Journal Information
Publication: Agriculture
Title: Unmanned Aerial Vehicle-Based Techniques for Monitoring and Prevention of Invasive Apple Snails (Pomacea canaliculata) in Rice Paddy Fields
DOI: 10.3390/agriculture14020299